Commit f3bb634348e974d5b52af6b79dfab2537e52c2f2 - netcdf4-python

[ Bas Couwenberg ] [ Debian Janitor ] New upstream snapshot. Debian Janitor 2 years ago

35 changed file(s) with 2574 addition(s) and 4922 deletion(s). Raw diff Collapse all Expand all

-95

~~.github/workflows/build.yml~~ less more

0		name: Build and Test Linux
1		on: [push, pull_request]
2		jobs:
3		build-linux:
4		name: Python (${{ matrix.python-version }})
5		runs-on: ubuntu-latest
6		env:
7		PNETCDF_VERSION: 1.12.1
8		NETCDF_VERSION: 4.7.4
9		NETCDF_DIR: ${{ github.workspace }}/..
10		NETCDF_EXTRA_CONFIG: --enable-pnetcdf
11		CC: mpicc.mpich
12		# NO_NET: 1
13		strategy:
14		matrix:
15		python-version: ["3.9"]
16		steps:
17
18		- uses: actions/checkout@v2
19
20		- name: Set up Python ${{ matrix.python-version }}
21		uses: actions/setup-python@v2
22		with:
23		python-version: ${{ matrix.python-version }}
24
25		- name: Install Ubuntu Dependencies
26		run: \|
27		sudo apt-get update
28		sudo apt-get install mpich libmpich-dev libhdf5-mpich-dev libcurl4-openssl-dev
29		echo "Download and build PnetCDF version ${PNETCDF_VERSION}"
30		wget https://parallel-netcdf.github.io/Release/pnetcdf-${PNETCDF_VERSION}.tar.gz
31		tar -xzf pnetcdf-${PNETCDF_VERSION}.tar.gz
32		pushd pnetcdf-${PNETCDF_VERSION}
33		./configure --prefix $NETCDF_DIR --enable-shared --disable-fortran --disable-cxx
34		make -j 2
35		make install
36		popd
37		echo "Download and build netCDF version ${NETCDF_VERSION}"
38		wget ftp://ftp.unidata.ucar.edu/pub/netcdf/netcdf-c-${NETCDF_VERSION}.tar.gz
39		tar -xzf netcdf-c-${NETCDF_VERSION}.tar.gz
40		pushd netcdf-c-${NETCDF_VERSION}
41		export CPPFLAGS="-I/usr/include/hdf5/mpich -I${NETCDF_DIR}/include"
42		export LDFLAGS="-L${NETCDF_DIR}/lib"
43		export LIBS="-lhdf5_mpich_hl -lhdf5_mpich -lm -lz"
44		./configure --prefix $NETCDF_DIR --enable-netcdf-4 --enable-shared --enable-dap --enable-parallel4 $NETCDF_EXTRA_CONFIG
45		make -j 2
46		make install
47		popd
48
49		# - name: The job has failed
50		# if: ${{ failure() }}
51		# run: \|
52		# cd netcdf-c-${NETCDF_VERSION}
53		# cat config.log
54
55		- name: Install python dependencies via pip
56		run: \|
57		python -m pip install --upgrade pip
58		pip install numpy cython cftime pytest twine wheel check-manifest mpi4py
59
60		- name: Install netcdf4-python
61		run: \|
62		export PATH=${NETCDF_DIR}/bin:${PATH}
63		python setup.py install
64		- name: Test
65		run: \|
66		export PATH=${NETCDF_DIR}/bin:${PATH}
67		python checkversion.py
68		# serial
69		cd test
70		python run_all.py
71		# parallel (hdf5 for netcdf4, pnetcdf for netcdf3)
72		cd ../examples
73		mpirun.mpich -np 4 python mpi_example.py
74		if [ $? -ne 0 ] ; then
75		echo "hdf5 mpi test failed!"
76		exit 1
77		else
78		echo "hdf5 mpi test passed!"
79		fi
80		mpirun.mpich -np 4 python mpi_example.py NETCDF3_64BIT_DATA
81		if [ $? -ne 0 ] ; then
82		echo "pnetcdf mpi test failed!"
83		exit 1
84		else
85		echo "pnetcdf mpi test passed!"
86		fi
87
88		- name: Tarball
89		run: \|
90		export PATH=${NETCDF_DIR}/bin:${PATH}
91		python setup.py --version
92		pip wheel . -w dist --no-deps
93		check-manifest --verbose
94		twine check dist/*

-85

~~.github/workflows/miniconda.yml~~ less more

0		name: Build and Test
1
2		on:
3		pull_request:
4		push:
5		branches: [master]
6
7		jobs:
8		run-serial:
9		runs-on: ${{ matrix.os }}
10		strategy:
11		matrix:
12		python-version: ["3.6", "3.7", "3.8", "3.9"]
13		os: [windows-latest, ubuntu-latest, macos-latest]
14		platform: [x64, x32]
15		exclude:
16		- os: macos-latest
17		platform: x32
18		steps:
19		- uses: actions/checkout@v2
20
21		- name: Setup Conda
22		uses: s-weigand/setup-conda@v1
23		with:
24		activate-conda: false
25		conda-channels: conda-forge
26
27		- name: Python ${{ matrix.python-version }}
28		shell: bash -l {0}
29		run: \|
30		conda create --name TEST python=${{ matrix.python-version }} numpy cython pip pytest hdf5 libnetcdf cftime --strict-channel-priority
31		source activate TEST
32		export PATH="${CONDA_PREFIX}/bin:${CONDA_PREFIX}/Library/bin:$PATH" # so setup.py finds nc-config
33		pip install -e . --no-deps --force-reinstall
34		conda info --all
35		conda list
36
37		- name: Tests
38		shell: bash -l {0}
39		run: \|
40		source activate TEST
41		cd test && python run_all.py
42
43		run-mpi:
44		runs-on: ${{ matrix.os }}
45		strategy:
46		matrix:
47		python-version: [ "3.9" ]
48		os: [ubuntu-latest]
49		platform: [x64]
50		steps:
51		- uses: actions/checkout@v2
52
53		- name: Setup Conda
54		uses: s-weigand/setup-conda@v1
55		with:
56		activate-conda: false
57		conda-channels: conda-forge
58
59		- name: Python ${{ matrix.python-version }}
60		shell: bash -l {0}
61		run: \|
62		conda create --name TEST python=${{ matrix.python-version }} numpy cython pip pytest mpi4py hdf5==mpi libnetcdf==mpi cftime --strict-channel-priority
63		source activate TEST
64		export PATH="${CONDA_PREFIX}/bin:${CONDA_PREFIX}/Library/bin:$PATH" # so setup.py finds nc-config
65		pip install -e . --no-deps --force-reinstall
66		conda info --all
67		conda list
68
69		- name: Tests
70		shell: bash -l {0}
71		run: \|
72		source activate TEST
73		cd test && python run_all.py
74		cd ../examples
75		export PATH="${CONDA_PREFIX}/bin:${CONDA_PREFIX}/Library/bin:$PATH"
76		which mpirun
77		mpirun --version
78		mpirun -np 4 python mpi_example.py
79		if [ $? -ne 0 ] ; then
80		echo "hdf5 mpi test failed!"
81		exit 1
82		else
83		echo "hdf5 mpi test passed!"
84		fi

-0

Changelog less more

	0	version 1.5.7 (not yet released)
	1	=================================
	2	* don't try to mask vlens with default _FillValue, since vlens don't have a default _FillValue.
	3	This gets rid of numpy DeprecationWarning (issue #1099).
	4	* update docs to reflect the fact that a variable must be in collective mode before writing
	5	compressed data to it in parallel. Added a test for this (examples/mpi_example_compressed.py).
	6	Issue #1108.
	7
0	8	version 1.5.6 (tag v1.5.6rel)
1	9	==============================
2	10	* move CI/CD tests from travis/appveyor to Github Actions (PR #1061).

+24

-0

PKG-INFO less more

	0	Metadata-Version: 1.1
	1	Name: netCDF4
	2	Version: 1.5.7
	3	Summary: Provides an object-oriented python interface to the netCDF version 4 library.
	4	Home-page: http://github.com/Unidata/netcdf4-python
	5	Author: Jeff Whitaker
	6	Author-email: jeffrey.s.whitaker@noaa.gov
	7	License: License :: OSI Approved :: MIT License
	8	Download-URL: http://python.org/pypi/netCDF4
	9	Description: netCDF version 4 has many features not found in earlier versions of the library, such as hierarchical groups, zlib compression, multiple unlimited dimensions, and new data types. It is implemented on top of HDF5. This module implements most of the new features, and can read and write netCDF files compatible with older versions of the library. The API is modelled after Scientific.IO.NetCDF, and should be familiar to users of that module.
	10
	11	This project is hosted on a `GitHub repository <https://github.com/Unidata/netcdf4-python>`_ where you may access the most up-to-date source.
	12	Keywords: numpy,netcdf,data,science,network,oceanography,meteorology,climate
	13	Platform: any
	14	Classifier: Development Status :: 3 - Alpha
	15	Classifier: Programming Language :: Python :: 3
	16	Classifier: Programming Language :: Python :: 3.6
	17	Classifier: Programming Language :: Python :: 3.7
	18	Classifier: Programming Language :: Python :: 3.8
	19	Classifier: Intended Audience :: Science/Research
	20	Classifier: License :: OSI Approved :: MIT License
	21	Classifier: Topic :: Software Development :: Libraries :: Python Modules
	22	Classifier: Topic :: System :: Archiving :: Compression
	23	Classifier: Operating System :: OS Independent

+23

-9

README.md less more

5	5	[![Anaconda-Server Badge](https://anaconda.org/conda-forge/netCDF4/badges/version.svg)](https://anaconda.org/conda-forge/netCDF4)
6	6	[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.2592291.svg)](https://doi.org/10.5281/zenodo.2592290)
7	7
	8
8	9	## News
9	10	For details on the latest updates, see the [Changelog](https://github.com/Unidata/netcdf4-python/blob/master/Changelog).
10	11
11		2/15/2020: Version [1.5.6](https://pypi.python.org/pypi/netCDF4/1.5.6) released. Added `Dataset.fromcdl` and `Dataset.tocdl`, which require `ncdump` and `ncgen` utilities to be in `$PATH`. Removed python 2.7 support.
	12	2/15/2021: Version [1.5.6](https://pypi.python.org/pypi/netCDF4/1.5.6) released. Added `Dataset.fromcdl` and `Dataset.tocdl`, which require `ncdump` and `ncgen` utilities to be in `$PATH`. Removed python 2.7 support.
12	13
13	14	12/20/2020: Version [1.5.5.1](https://pypi.python.org/pypi/netCDF4/1.5.5.1) released.
14	15	Updated binary wheels for OSX and linux that link latest netcdf-c and hdf5 libs.

205	206	with `export USE_NCCONFIG=0` will use environment variables to find paths to libraries and include files,
206	207	instead of relying exclusively on the nc-config utility.
207	208
208		## Quick Start
209		* Clone GitHub repository (`git clone https://github.com/Unidata/netcdf4-python.git`), or get source tarball from [PyPI](https://pypi.python.org/pypi/netCDF4). Links to Windows and OS X precompiled binary packages are also available on [PyPI](https://pypi.python.org/pypi/netCDF4).
	209	## Installation
	210	The easiest way to install is through pip:
	211
	212	```shell
	213	pip install netCDF4
	214	```
	215
	216	or, if you are a user of the Conda package manager,
	217
	218	```shell
	219	conda install -c conda-forge netCDF4
	220	```
	221
	222	## Development installation
	223	* Clone GitHub repository (`git clone https://github.com/Unidata/netcdf4-python.git`)
210	224
211	225	* Make sure [numpy](http://www.numpy.org/) and [Cython](http://cython.org/) are
212		installed and you have [Python](https://www.python.org) 2.7 or newer.
213
214		* Make sure [HDF5](http://www.h5py.org/) and netcdf-4 are installed, and the `nc-config` utility
215		is in your Unix PATH.
216
217		* Run `python setup.py build`, then `python setup.py install` (with `sudo` if necessary).
	226	installed and you have [Python](https://www.python.org) 3.6 or newer.
	227
	228	* Make sure [HDF5](http://www.h5py.org/) and netcdf-4 are installed,
	229	and the `nc-config` utility is in your Unix PATH.
	230
	231	* Run `python setup.py build`, then `pip install -e .`.
218	232
219	233	* To run all the tests, execute `cd test && python run_all.py`.
220	234

-2

create_docs.sh less more

0	0	# Uses pdoc (https://github.com/mitmproxy/pdoc)
1	1	# to create html docs from docstrings in Cython source.
2		/Users/jwhitaker/.local/bin/pdoc -o 'docs' netCDF4
	2	pdoc -o 'docs' netCDF4
3	3	# use resulting docs/netCDF4/_netCDF4.html
4		cp docs/netCDF4/_netCDF4.html docs/index.html
	4	cp docs/netCDF4.html docs/index.html
5	5	sed -i -e 's!href="../netCDF4.html!href="./index.html!g' docs/index.html
6	6	sed -i -e 's!/../netCDF4.html!/index.html!g' docs/index.html
7	7	sed -i -e 's!._netCDF4 API! API!g' docs/index.html

-3

debian/changelog less more

0		netcdf4-python (1.5.6-1~exp2) UNRELEASED; urgency=medium
1
	0	netcdf4-python (1.5.6rel+git20210522.1.203dcc1-1) UNRELEASED; urgency=medium
	1
	2	[ Bas Couwenberg ]
2	3	* Update watch file for GitHub URL changes.
3	4
4		-- Bas Couwenberg <sebastic@debian.org> Sat, 20 Mar 2021 08:45:38 +0100
	5	[ Debian Janitor ]
	6	* New upstream snapshot.
	7
	8	-- Bas Couwenberg <sebastic@debian.org> Wed, 02 Jun 2021 05:04:08 -0000
5	9
6	10	netcdf4-python (1.5.6-1~exp1) experimental; urgency=medium
7	11

+1942

-1737

docs/index.html less more

2	2	<head>
3	3	<meta charset="utf-8">
4	4	<meta name="viewport" content="width=device-width, initial-scale=1">
5		<meta name="generator" content="pdoc" />
	5	<meta name="generator" content="pdoc 6.4.3" />
6	6	<title>netCDF4 API documentation</title>
7	7	<link rel="icon" type="image/svg+xml" href="data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A//www.w3.org/2000/svg%22%20width%3D%2264%22%20height%3D%2264%22%20viewBox%3D%2244.5%202.5%2015%2015%22%3E%3Cpath%20d%3D%22M49.351%2021.041c-.233-.721-.546-2.408-.772-4.076-.042-.09-.067-.187-.046-.288-.166-1.347-.277-2.625-.241-3.351-1.378-1.008-2.271-2.586-2.271-4.362%200-.976.272-1.935.788-2.774.057-.094.122-.18.184-.268-.033-.167-.052-.339-.052-.516%200-1.477%201.202-2.679%202.679-2.679.791%200%201.496.352%201.987.9a6.3%206.3%200%200%201%201.001.029c.492-.564%201.207-.929%202.012-.929%201.477%200%202.679%201.202%202.679%202.679a2.65%202.65%200%200%201-.269%201.148c.383.747.595%201.572.595%202.41%200%202.311-1.507%204.29-3.635%205.107.037.699.147%202.27.423%203.294l.137.461c.156%202.136-4.612%205.166-5.199%203.215zm.127-4.919a4.78%204.78%200%200%200%20.775-.584c-.172-.115-.505-.254-.88-.378zm.331%202.302l.828-.502c-.202-.143-.576-.328-.984-.49zm.45%202.157l.701-.403c-.214-.115-.536-.249-.891-.376l.19.779zM49.13%204.141c0%20.152.123.276.276.276s.275-.124.275-.276-.123-.276-.276-.276-.275.124-.275.276zm.735-.389a1.15%201.15%200%200%201%20.314.783%201.16%201.16%200%200%201-1.162%201.162c-.457%200-.842-.27-1.032-.653-.026.117-.042.238-.042.362a1.68%201.68%200%200%200%201.679%201.679%201.68%201.68%200%200%200%201.679-1.679c0-.843-.626-1.535-1.436-1.654zm3.076%201.654a1.68%201.68%200%200%200%201.679%201.679%201.68%201.68%200%200%200%201.679-1.679c0-.037-.009-.072-.011-.109-.21.3-.541.508-.935.508a1.16%201.16%200%200%201-1.162-1.162%201.14%201.14%200%200%201%20.474-.912c-.015%200-.03-.005-.045-.005-.926.001-1.679.754-1.679%201.68zm1.861-1.265c0%20.152.123.276.276.276s.275-.124.275-.276-.123-.276-.276-.276-.275.124-.275.276zm1.823%204.823c0-.52-.103-1.035-.288-1.52-.466.394-1.06.64-1.717.64-1.144%200-2.116-.725-2.499-1.738-.383%201.012-1.355%201.738-2.499%201.738-.867%200-1.631-.421-2.121-1.062-.307.605-.478%201.267-.478%201.942%200%202.486%202.153%204.51%204.801%204.51s4.801-2.023%204.801-4.51zm-3.032%209.156l-.146-.492c-.276-1.02-.395-2.457-.444-3.268a6.11%206.11%200%200%201-1.18.115%206.01%206.01%200%200%201-2.536-.562l.006.175c.802.215%201.848.612%202.021%201.25.079.295-.021.601-.274.837l-.598.501c.667.304%201.243.698%201.311%201.179.02.144.022.507-.393.787l-.564.365c1.285.521%201.361.96%201.381%201.126.018.142.011.496-.427.746l-.854.489c.064-1.19%201.985-2.585%202.697-3.248zM49.34%209.925c0-.667%201-.667%201%200%200%20.653.818%201.205%201.787%201.205s1.787-.552%201.787-1.205c0-.667%201-.667%201%200%200%201.216-1.25%202.205-2.787%202.205s-2.787-.989-2.787-2.205zm-.887-7.633c-.093.077-.205.114-.317.114a.5.5%200%200%201-.318-.886L49.183.397a.5.5%200%200%201%20.703.068.5.5%200%200%201-.069.703zm7.661-.065c-.086%200-.173-.022-.253-.068l-1.523-.893c-.575-.337-.069-1.2.506-.863l1.523.892a.5.5%200%200%201%20.179.685c-.094.158-.261.247-.432.247z%22%20fill%3D%22%233bb300%22/%3E%3C/svg%3E"/>
8	8
9	9
10	10	<style type="text/css">/! Bootstrap Reboot v5.0.0-beta1 (https://getbootstrap.com/) * Copyright 2011-2020 The Bootstrap Authors * Copyright 2011-2020 Twitter, Inc. * Licensed under MIT (https://github.com/twbs/bootstrap/blob/main/LICENSE) * Forked from Normalize.css, licensed MIT (https://github.com/necolas/normalize.css/blob/master/LICENSE.md) /,::after,::before{box-sizing:border-box}@media (prefers-reduced-motion:no-preference){:root{scroll-behavior:smooth}}body{margin:0;font-family:system-ui,-apple-system,"Segoe UI",Roboto,"Helvetica Neue",Arial,"Noto Sans","Liberation Sans",sans-serif,"Apple Color Emoji","Segoe UI Emoji","Segoe UI Symbol","Noto Color Emoji";font-size:1rem;font-weight:400;line-height:1.5;color:#212529;background-color:#fff;-webkit-text-size-adjust:100%;-webkit-tap-highlight-color:transparent}[tabindex="-1"]:focus:not(:focus-visible){outline:0!important}hr{margin:1rem 0;color:inherit;background-color:currentColor;border:0;opacity:.25}hr:not([size]){height:1px}h1,h2,h3,h4,h5,h6{margin-top:0;margin-bottom:.5rem;font-weight:500;line-height:1.2}h1{font-size:calc(1.375rem + 1.5vw)}@media (min-width:1200px){h1{font-size:2.5rem}}h2{font-size:calc(1.325rem + .9vw)}@media (min-width:1200px){h2{font-size:2rem}}h3{font-size:calc(1.3rem + .6vw)}@media (min-width:1200px){h3{font-size:1.75rem}}h4{font-size:calc(1.275rem + .3vw)}@media (min-width:1200px){h4{font-size:1.5rem}}h5{font-size:1.25rem}h6{font-size:1rem}p{margin-top:0;margin-bottom:1rem}abbr[data-bs-original-title],abbr[title]{text-decoration:underline;-webkit-text-decoration:underline dotted;text-decoration:underline dotted;cursor:help;-webkit-text-decoration-skip-ink:none;text-decoration-skip-ink:none}address{margin-bottom:1rem;font-style:normal;line-height:inherit}ol,ul{padding-left:2rem}dl,ol,ul{margin-top:0;margin-bottom:1rem}ol ol,ol ul,ul ol,ul ul{margin-bottom:0}dt{font-weight:700}dd{margin-bottom:.5rem;margin-left:0}blockquote{margin:0 0 1rem}b,strong{font-weight:bolder}small{font-size:.875em}mark{padding:.2em;background-color:#fcf8e3}sub,sup{position:relative;font-size:.75em;line-height:0;vertical-align:baseline}sub{bottom:-.25em}sup{top:-.5em}a{color:#0d6efd;text-decoration:underline}a:hover{color:#0a58ca}a:not([href]):not([class]),a:not([href]):not([class]):hover{color:inherit;text-decoration:none}code,kbd,pre,samp{font-family:SFMono-Regular,Menlo,Monaco,Consolas,"Liberation Mono","Courier New",monospace;font-size:1em;direction:ltr;unicode-bidi:bidi-override}pre{display:block;margin-top:0;margin-bottom:1rem;overflow:auto;font-size:.875em}pre code{font-size:inherit;color:inherit;word-break:normal}code{font-size:.875em;color:#d63384;word-wrap:break-word}a>code{color:inherit}kbd{padding:.2rem .4rem;font-size:.875em;color:#fff;background-color:#212529;border-radius:.2rem}kbd kbd{padding:0;font-size:1em;font-weight:700}figure{margin:0 0 1rem}img,svg{vertical-align:middle}table{caption-side:bottom;border-collapse:collapse}caption{padding-top:.5rem;padding-bottom:.5rem;color:#6c757d;text-align:left}th{text-align:inherit;text-align:-webkit-match-parent}tbody,td,tfoot,th,thead,tr{border-color:inherit;border-style:solid;border-width:0}label{display:inline-block}button{border-radius:0}button:focus{outline:dotted 1px;outline:-webkit-focus-ring-color auto 5px}button,input,optgroup,select,textarea{margin:0;font-family:inherit;font-size:inherit;line-height:inherit}button,select{text-transform:none}[role=button]{cursor:pointer}select{word-wrap:normal}[list]::-webkit-calendar-picker-indicator{display:none}[type=button],[type=reset],[type=submit],button{-webkit-appearance:button}[type=button]:not(:disabled),[type=reset]:not(:disabled),[type=submit]:not(:disabled),button:not(:disabled){cursor:pointer}::-moz-focus-inner{padding:0;border-style:none}textarea{resize:vertical}fieldset{min-width:0;padding:0;margin:0;border:0}legend{float:left;width:100%;padding:0;margin-bottom:.5rem;font-size:calc(1.275rem + .3vw);line-height:inherit}@media (min-width:1200px){legend{font-size:1.5rem}}legend+*{clear:left}::-webkit-datetime-edit-day-field,::-webkit-datetime-edit-fields-wrapper,::-webkit-datetime-edit-hour-field,::-webkit-datetime-edit-minute,::-webkit-datetime-edit-month-field,::-webkit-datetime-edit-text,::-webkit-datetime-edit-year-field{padding:0}::-webkit-inner-spin-button{height:auto}[type=search]{outline-offset:-2px;-webkit-appearance:textfield}::-webkit-search-decoration{-webkit-appearance:none}::-webkit-color-swatch-wrapper{padding:0}::file-selector-button{font:inherit}::-webkit-file-upload-button{font:inherit;-webkit-appearance:button}output{display:inline-block}iframe{border:0}summary{display:list-item;cursor:pointer}progress{vertical-align:baseline}[hidden]{display:none!important}</style>
11	11	<style type="text/css">/! pygments syntax highlighting /pre{line-height:125%;}td.linenos pre{color:#000000; background-color:#f0f0f0; padding-left:5px; padding-right:5px;}span.linenos{color:#000000; background-color:#f0f0f0; padding-left:5px; padding-right:5px;}td.linenos pre.special{color:#000000; background-color:#ffffc0; padding-left:5px; padding-right:5px;}span.linenos.special{color:#000000; background-color:#ffffc0; padding-left:5px; padding-right:5px;}.pdoc .hll{background-color:#ffffcc}.pdoc{background:#f8f8f8;}.pdoc .c{color:#408080; font-style:italic}.pdoc .err{border:1px solid #FF0000}.pdoc .k{color:#008000; font-weight:bold}.pdoc .o{color:#666666}.pdoc .ch{color:#408080; font-style:italic}.pdoc .cm{color:#408080; font-style:italic}.pdoc .cp{color:#BC7A00}.pdoc .cpf{color:#408080; font-style:italic}.pdoc .c1{color:#408080; font-style:italic}.pdoc .cs{color:#408080; font-style:italic}.pdoc .gd{color:#A00000}.pdoc .ge{font-style:italic}.pdoc .gr{color:#FF0000}.pdoc .gh{color:#000080; font-weight:bold}.pdoc .gi{color:#00A000}.pdoc .go{color:#888888}.pdoc .gp{color:#000080; font-weight:bold}.pdoc .gs{font-weight:bold}.pdoc .gu{color:#800080; font-weight:bold}.pdoc .gt{color:#0044DD}.pdoc .kc{color:#008000; font-weight:bold}.pdoc .kd{color:#008000; font-weight:bold}.pdoc .kn{color:#008000; font-weight:bold}.pdoc .kp{color:#008000}.pdoc .kr{color:#008000; font-weight:bold}.pdoc .kt{color:#B00040}.pdoc .m{color:#666666}.pdoc .s{color:#BA2121}.pdoc .na{color:#7D9029}.pdoc .nb{color:#008000}.pdoc .nc{color:#0000FF; font-weight:bold}.pdoc .no{color:#880000}.pdoc .nd{color:#AA22FF}.pdoc .ni{color:#999999; font-weight:bold}.pdoc .ne{color:#D2413A; font-weight:bold}.pdoc .nf{color:#0000FF}.pdoc .nl{color:#A0A000}.pdoc .nn{color:#0000FF; font-weight:bold}.pdoc .nt{color:#008000; font-weight:bold}.pdoc .nv{color:#19177C}.pdoc .ow{color:#AA22FF; font-weight:bold}.pdoc .w{color:#bbbbbb}.pdoc .mb{color:#666666}.pdoc .mf{color:#666666}.pdoc .mh{color:#666666}.pdoc .mi{color:#666666}.pdoc .mo{color:#666666}.pdoc .sa{color:#BA2121}.pdoc .sb{color:#BA2121}.pdoc .sc{color:#BA2121}.pdoc .dl{color:#BA2121}.pdoc .sd{color:#BA2121; font-style:italic}.pdoc .s2{color:#BA2121}.pdoc .se{color:#BB6622; font-weight:bold}.pdoc .sh{color:#BA2121}.pdoc .si{color:#BB6688; font-weight:bold}.pdoc .sx{color:#008000}.pdoc .sr{color:#BB6688}.pdoc .s1{color:#BA2121}.pdoc .ss{color:#19177C}.pdoc .bp{color:#008000}.pdoc .fm{color:#0000FF}.pdoc .vc{color:#19177C}.pdoc .vg{color:#19177C}.pdoc .vi{color:#19177C}.pdoc .vm{color:#19177C}.pdoc .il{color:#666666}</style>
12		<style type="text/css">/! pdoc /:root{--pdoc-background:#fff;}.pdoc{--text:#212529;--muted:#6c757d;--link:#3660a5;--link-hover:#1659c5;--code:#f7f7f7;--active:#fff598;--accent:#eee;--accent2:#c1c1c1;--nav-hover:rgba(255, 255, 255, 0.5);--name:#0066BB;--def:#008800;--annotation:#007020;}body{background-color:var(--pdoc-background);}html, body{width:100%;height:100%;}@media (max-width:769px){#navtoggle{cursor:pointer;position:absolute;width:50px;height:40px;top:1rem;right:1rem;border-color:var(--text);color:var(--text);display:flex;opacity:0.8;}#navtoggle:hover{opacity:1;}#togglestate + div{display:none;}#togglestate:checked + div{display:inherit;}main{padding:2rem 3vw;}.git-button{display:none !important;}}@media (min-width:770px){:root{--sidebar-width:clamp(12.5rem, 28vw, 22rem);}nav{position:fixed;overflow:auto;height:100vh;width:var(--sidebar-width);}main{padding:3rem 2rem 3rem calc(var(--sidebar-width) + 3rem);width:calc(54rem + var(--sidebar-width));max-width:100%;}#navtoggle{display:none;}}#togglestate{display:none;}nav.pdoc{--pad:1.75rem;--indent:1.5rem;background-color:var(--accent);border-right:1px solid var(--accent2);box-shadow:0 0 20px rgba(50, 50, 50, .2) inset;padding:0 0 0 var(--pad);overflow-wrap:anywhere;scrollbar-width:thin; scrollbar-color:var(--accent2) transparent }nav.pdoc::-webkit-scrollbar{width:.4rem; }nav.pdoc::-webkit-scrollbar-thumb{background-color:var(--accent2); }nav.pdoc > div{padding:var(--pad) 0;}nav.pdoc .module-list-button{display:inline-flex;align-items:center;color:var(--text);border-color:var(--muted);margin-bottom:1rem;}nav.pdoc .module-list-button:hover{border-color:var(--text);}nav.pdoc ul{list-style:none;padding-left:0;}nav.pdoc li{display:block;margin:0;padding:.2rem 0 .2rem var(--indent);transition:all 100ms;}nav.pdoc > div > ul > li{padding-left:0;}nav.pdoc li:hover{background-color:var(--nav-hover);}nav.pdoc a, nav.pdoc a:hover{color:var(--text);}nav.pdoc a{display:block;}nav.pdoc > h2:first-of-type{margin-top:1.5rem;}nav.pdoc .class:before{content:"class ";color:var(--muted);}nav.pdoc .function:after{content:"()";color:var(--muted);}html, main{scroll-behavior:smooth;}.pdoc{color:var(--text);box-sizing:border-box;line-height:1.5;background:none;}.pdoc .pdoc-button{display:inline-block;border:solid black 1px;border-radius:2px;font-size:.75rem;padding:calc(0.5em - 1px) 1em;transition:100ms all;}.pdoc h1, .pdoc h2, .pdoc h3, .pdoc h4{font-weight:300;margin:.3em 0;padding:.2em 0;}.pdoc a{text-decoration:none;color:var(--link);}.pdoc a:hover{color:var(--link-hover);}.pdoc blockquote{margin-left:2rem;}.pdoc pre{background-color:var(--code);border-top:1px solid var(--accent2);border-bottom:1px solid var(--accent2);margin-bottom:1em;padding:.5rem 0 .5rem .5rem;overflow-x:auto;}.pdoc code{color:var(--text);padding:.2em .4em;margin:0;font-size:85%;background-color:var(--code);border-radius:6px;}.pdoc a > code{color:inherit;}.pdoc pre > code{display:inline-block;font-size:inherit;background:none;border:none;padding:0;}.pdoc .modulename{margin-top:0;font-weight:bold;}.pdoc .modulename a{color:var(--link);transition:100ms all;}.pdoc .git-button{float:right;border:solid var(--link) 1px;}.pdoc .git-button:hover{background-color:var(--link);color:var(--pdoc-background);}.pdoc details{--shift:-40px;text-align:right;margin-top:var(--shift);margin-bottom:calc(0px - var(--shift));clear:both;filter:opacity(1);}.pdoc details:not([open]){height:0;overflow:visible;}.pdoc details > summary{font-size:.75rem;cursor:pointer;color:var(--muted);border-width:0;padding:0 .7em;display:inline-block;display:inline list-item;user-select:none;}.pdoc details > summary:focus{outline:0;}.pdoc details > div{margin-top:calc(0px - var(--shift) / 2);text-align:left;}.pdoc .docstring{margin-bottom:1.5rem;}.pdoc > section:first-of-type > .docstring{margin-bottom:3rem;}.pdoc .docstring pre{margin-left:1em;margin-right:1em;}.pdoc h1:target,.pdoc h2:target,.pdoc h3:target,.pdoc h4:target,.pdoc h5:target,.pdoc h6:target{background-color:var(--active);box-shadow:-1rem 0 0 0 var(--active);}.pdoc div:target > .attr,.pdoc section:target > .attr,.pdoc dd:target > a{background-color:var(--active);}.pdoc .attr:hover{filter:contrast(0.95);}.pdoc .headerlink{position:absolute;width:0;margin-left:-1.5rem;line-height:1.4rem;font-size:1.5rem;font-weight:normal;transition:all 100ms ease-in-out;opacity:0;}.pdoc .attr > .headerlink{margin-left:-2.5rem;}.pdoc :hover > .headerlink,.pdoc :target > .attr > .headerlink{opacity:1;}.pdoc .attr{color:var(--text);margin:1rem 0 .5rem;padding:.4rem 5rem .4rem 1rem;background-color:var(--accent);}.pdoc .classattr{margin-left:2rem;}.pdoc .name{color:var(--name);font-weight:bold;}.pdoc .def{color:var(--def);font-weight:bold;}.pdoc .signature{white-space:pre-wrap;}.pdoc .annotation{color:var(--annotation);}.pdoc .inherited{margin-left:2rem;}.pdoc .inherited dt{font-weight:700;}.pdoc .inherited dt, .pdoc .inherited dd{display:inline;margin-left:0;margin-bottom:.5rem;}.pdoc .inherited dd:not(:last-child):after{content:", ";}.pdoc .inherited .class:before{content:"class ";}.pdoc .inherited .function a:after{content:"()";}.pdoc .attribution{margin-top:2rem;display:block;opacity:0.5;transition:all 200ms;filter:grayscale(100%);}.pdoc .attribution:hover{opacity:1;filter:grayscale(0%);}.pdoc .attribution img{margin-left:5px;height:35px;vertical-align:middle;width:70px;transition:all 200ms;}</style>
	12	<style type="text/css">/! pdoc /:root{--pdoc-background:#fff;}.pdoc{--text:#212529;--muted:#6c757d;--link:#3660a5;--link-hover:#1659c5;--code:#f7f7f7;--active:#fff598;--accent:#eee;--accent2:#c1c1c1;--nav-hover:rgba(255, 255, 255, 0.5);--name:#0066BB;--def:#008800;--annotation:#007020;}body{background-color:var(--pdoc-background);}html, body{width:100%;height:100%;}@media (max-width:769px){#navtoggle{cursor:pointer;position:absolute;width:50px;height:40px;top:1rem;right:1rem;border-color:var(--text);color:var(--text);display:flex;opacity:0.8;}#navtoggle:hover{opacity:1;}#togglestate + div{display:none;}#togglestate:checked + div{display:inherit;}main{padding:2rem 3vw;}.git-button{display:none !important;}}@media (min-width:770px){:root{--sidebar-width:clamp(12.5rem, 28vw, 22rem);}nav{position:fixed;overflow:auto;height:100vh;width:var(--sidebar-width);}main{padding:3rem 2rem 3rem calc(var(--sidebar-width) + 3rem);width:calc(54rem + var(--sidebar-width));max-width:100%;}#navtoggle{display:none;}}#togglestate{display:none;}nav.pdoc{--pad:1.75rem;--indent:1.5rem;background-color:var(--accent);border-right:1px solid var(--accent2);box-shadow:0 0 20px rgba(50, 50, 50, .2) inset;padding:0 0 0 var(--pad);overflow-wrap:anywhere;scrollbar-width:thin; scrollbar-color:var(--accent2) transparent }nav.pdoc::-webkit-scrollbar{width:.4rem; }nav.pdoc::-webkit-scrollbar-thumb{background-color:var(--accent2); }nav.pdoc > div{padding:var(--pad) 0;}nav.pdoc .module-list-button{display:inline-flex;align-items:center;color:var(--text);border-color:var(--muted);margin-bottom:1rem;}nav.pdoc .module-list-button:hover{border-color:var(--text);}nav.pdoc ul{list-style:none;padding-left:0;}nav.pdoc li{display:block;margin:0;padding:.2rem 0 .2rem var(--indent);transition:all 100ms;}nav.pdoc > div > ul > li{padding-left:0;}nav.pdoc li:hover{background-color:var(--nav-hover);}nav.pdoc a, nav.pdoc a:hover{color:var(--text);}nav.pdoc a{display:block;}nav.pdoc > h2:first-of-type{margin-top:1.5rem;}nav.pdoc .class:before{content:"class ";color:var(--muted);}nav.pdoc .function:after{content:"()";color:var(--muted);}html, main{scroll-behavior:smooth;}.pdoc{color:var(--text);box-sizing:border-box;line-height:1.5;background:none;}.pdoc .pdoc-button{display:inline-block;border:solid black 1px;border-radius:2px;font-size:.75rem;padding:calc(0.5em - 1px) 1em;transition:100ms all;}.pdoc .visually-hidden{position:absolute !important;width:1px !important;height:1px !important;padding:0 !important;margin:-1px !important;overflow:hidden !important;clip:rect(0, 0, 0, 0) !important;white-space:nowrap !important;border:0 !important;}.pdoc h1, .pdoc h2, .pdoc h3, .pdoc h4{font-weight:300;margin:.3em 0;padding:.2em 0;}.pdoc a{text-decoration:none;color:var(--link);}.pdoc a:hover{color:var(--link-hover);}.pdoc blockquote{margin-left:2rem;}.pdoc pre{background-color:var(--code);border-top:1px solid var(--accent2);border-bottom:1px solid var(--accent2);margin-bottom:1em;padding:.5rem 0 .5rem .5rem;overflow-x:auto;}.pdoc code{color:var(--text);padding:.2em .4em;margin:0;font-size:85%;background-color:var(--code);border-radius:6px;}.pdoc a > code{color:inherit;}.pdoc pre > code{display:inline-block;font-size:inherit;background:none;border:none;padding:0;}.pdoc .modulename{margin-top:0;font-weight:bold;}.pdoc .modulename a{color:var(--link);transition:100ms all;}.pdoc .git-button{float:right;border:solid var(--link) 1px;}.pdoc .git-button:hover{background-color:var(--link);color:var(--pdoc-background);}.pdoc details{--shift:-40px;text-align:right;margin-top:var(--shift);margin-bottom:calc(0px - var(--shift));clear:both;filter:opacity(1);}.pdoc details:not([open]){height:0;overflow:visible;}.pdoc details > summary{font-size:.75rem;cursor:pointer;color:var(--muted);border-width:0;padding:0 .7em;display:inline-block;display:inline list-item;user-select:none;}.pdoc details > summary:focus{outline:0;}.pdoc details > div{margin-top:calc(0px - var(--shift) / 2);text-align:left;}.pdoc .docstring{margin-bottom:1.5rem;}.pdoc > section:first-of-type > .docstring{margin-bottom:3rem;}.pdoc .docstring pre{margin-left:1em;margin-right:1em;}.pdoc h1:target,.pdoc h2:target,.pdoc h3:target,.pdoc h4:target,.pdoc h5:target,.pdoc h6:target{background-color:var(--active);box-shadow:-1rem 0 0 0 var(--active);}.pdoc div:target > .attr,.pdoc section:target > .attr,.pdoc dd:target > a{background-color:var(--active);}.pdoc .attr:hover{filter:contrast(0.95);}.pdoc .headerlink{position:absolute;width:0;margin-left:-1.5rem;line-height:1.4rem;font-size:1.5rem;font-weight:normal;transition:all 100ms ease-in-out;opacity:0;}.pdoc .attr > .headerlink{margin-left:-2.5rem;}.pdoc :hover > .headerlink,.pdoc :target > .attr > .headerlink{opacity:1;}.pdoc .attr{color:var(--text);margin:1rem 0 .5rem;padding:.4rem 5rem .4rem 1rem;background-color:var(--accent);}.pdoc .classattr{margin-left:2rem;}.pdoc .name{color:var(--name);font-weight:bold;}.pdoc .def{color:var(--def);font-weight:bold;}.pdoc .signature{white-space:pre-wrap;}.pdoc .annotation{color:var(--annotation);}.pdoc .inherited{margin-left:2rem;}.pdoc .inherited dt{font-weight:700;}.pdoc .inherited dt, .pdoc .inherited dd{display:inline;margin-left:0;margin-bottom:.5rem;}.pdoc .inherited dd:not(:last-child):after{content:", ";}.pdoc .inherited .class:before{content:"class ";}.pdoc .inherited .function a:after{content:"()";}.pdoc .attribution{margin-top:2rem;display:block;opacity:0.5;transition:all 200ms;filter:grayscale(100%);}.pdoc .attribution:hover{opacity:1;filter:grayscale(0%);}.pdoc .attribution img{margin-left:5px;height:35px;vertical-align:middle;width:70px;transition:all 200ms;}</style>
13	13	</head>
14	14	<body> <nav class="pdoc">
15	15	<label id="navtoggle" for="togglestate" class="pdoc-button"><svg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 30 30'><path stroke-linecap='round' stroke="currentColor" stroke-miterlimit='10' stroke-width='2' d='M4 7h22M4 15h22M4 23h22'/></svg></label>
16	16	<input id="togglestate" type="checkbox">
17	17	<div>
18		<a class="pdoc-button module-list-button" href="./index.html">
19		<svg xmlns="http://www.w3.org/2000/svg" width="16" height="16" fill="currentColor" class="bi bi-box-arrow-in-left" viewBox="0 0 16 16">
20		<path fill-rule="evenodd" d="M10 3.5a.5.5 0 0 0-.5-.5h-8a.5.5 0 0 0-.5.5v9a.5.5 0 0 0 .5.5h8a.5.5 0 0 0 .5-.5v-2a.5.5 0 0 1 1 0v2A1.5 1.5 0 0 1 9.5 14h-8A1.5 1.5 0 0 1 0 12.5v-9A1.5 1.5 0 0 1 1.5 2h8A1.5 1.5 0 0 1 11 3.5v2a.5.5 0 0 1-1 0v-2z"/>
21		<path fill-rule="evenodd" d="M4.146 8.354a.5.5 0 0 1 0-.708l3-3a.5.5 0 1 1 .708.708L5.707 7.5H14.5a.5.5 0 0 1 0 1H5.707l2.147 2.146a.5.5 0 0 1-.708.708l-3-3z"/>
22		</svg>  netCDF4</a>
23	18
24	19
25	20	<h2>Contents</h2>
26	21	<ul>
27		<li><a href="#version-156">Version 1.5.6</a></li>
	22	<li><a href="#version-157">Version 1.5.7</a></li>
28	23	</ul></li>
29	24	<li><a href="#introduction">Introduction</a>
30	25	<ul>
31		<li><a href="#download">Download</a></li>
32		<li><a href="#requires">Requires</a></li>
33		<li><a href="#install">Install</a></li>
	26	<li><a href="#quick-install">Quick Install</a></li>
	27	<li><a href="#developer-install">Developer Install</a></li>
34	28	</ul></li>
35	29	<li><a href="#tutorial">Tutorial</a>
36	30	<ul>

197	191
198	192	</li>
199	193	<li>
200		<a class="class" href="#Group">Group</a>
	194	<a class="class" href="#Variable">Variable</a>
201	195	<ul class="memberlist">
202	196	<li>
203		<a class="function" href="#Group.__init__">Group</a>
204		</li>
205		<li>
206		<a class="function" href="#Group.close">close</a>
	197	<a class="function" href="#Variable.__init__">Variable</a>
	198	</li>
	199	<li>
	200	<a class="function" href="#Variable.group">group</a>
	201	</li>
	202	<li>
	203	<a class="function" href="#Variable.ncattrs">ncattrs</a>
	204	</li>
	205	<li>
	206	<a class="function" href="#Variable.setncattr">setncattr</a>
	207	</li>
	208	<li>
	209	<a class="function" href="#Variable.setncattr_string">setncattr_string</a>
	210	</li>
	211	<li>
	212	<a class="function" href="#Variable.setncatts">setncatts</a>
	213	</li>
	214	<li>
	215	<a class="function" href="#Variable.getncattr">getncattr</a>
	216	</li>
	217	<li>
	218	<a class="function" href="#Variable.delncattr">delncattr</a>
	219	</li>
	220	<li>
	221	<a class="function" href="#Variable.filters">filters</a>
	222	</li>
	223	<li>
	224	<a class="function" href="#Variable.endian">endian</a>
	225	</li>
	226	<li>
	227	<a class="function" href="#Variable.chunking">chunking</a>
	228	</li>
	229	<li>
	230	<a class="function" href="#Variable.get_var_chunk_cache">get_var_chunk_cache</a>
	231	</li>
	232	<li>
	233	<a class="function" href="#Variable.set_var_chunk_cache">set_var_chunk_cache</a>
	234	</li>
	235	<li>
	236	<a class="function" href="#Variable.renameAttribute">renameAttribute</a>
	237	</li>
	238	<li>
	239	<a class="function" href="#Variable.assignValue">assignValue</a>
	240	</li>
	241	<li>
	242	<a class="function" href="#Variable.getValue">getValue</a>
	243	</li>
	244	<li>
	245	<a class="function" href="#Variable.set_auto_chartostring">set_auto_chartostring</a>
	246	</li>
	247	<li>
	248	<a class="function" href="#Variable.use_nc_get_vars">use_nc_get_vars</a>
	249	</li>
	250	<li>
	251	<a class="function" href="#Variable.set_auto_maskandscale">set_auto_maskandscale</a>
	252	</li>
	253	<li>
	254	<a class="function" href="#Variable.set_auto_scale">set_auto_scale</a>
	255	</li>
	256	<li>
	257	<a class="function" href="#Variable.set_auto_mask">set_auto_mask</a>
	258	</li>
	259	<li>
	260	<a class="function" href="#Variable.set_always_mask">set_always_mask</a>
	261	</li>
	262	<li>
	263	<a class="function" href="#Variable.set_ncstring_attrs">set_ncstring_attrs</a>
	264	</li>
	265	<li>
	266	<a class="function" href="#Variable.set_collective">set_collective</a>
	267	</li>
	268	<li>
	269	<a class="function" href="#Variable.get_dims">get_dims</a>
	270	</li>
	271	<li>
	272	<a class="variable" href="#Variable.name">name</a>
	273	</li>
	274	<li>
	275	<a class="variable" href="#Variable.datatype">datatype</a>
	276	</li>
	277	<li>
	278	<a class="variable" href="#Variable.shape">shape</a>
	279	</li>
	280	<li>
	281	<a class="variable" href="#Variable.size">size</a>
	282	</li>
	283	<li>
	284	<a class="variable" href="#Variable.dimensions">dimensions</a>
	285	</li>
	286	<li>
	287	<a class="variable" href="#Variable.ndim">ndim</a>
	288	</li>
	289	<li>
	290	<a class="variable" href="#Variable.dtype">dtype</a>
	291	</li>
	292	<li>
	293	<a class="variable" href="#Variable.mask">mask</a>
	294	</li>
	295	<li>
	296	<a class="variable" href="#Variable.scale">scale</a>
	297	</li>
	298	<li>
	299	<a class="variable" href="#Variable.always_mask">always_mask</a>
	300	</li>
	301	<li>
	302	<a class="variable" href="#Variable.chartostring">chartostring</a>
207	303	</li>
208	304	</ul>
209	305

230	326
231	327	</li>
232	328	<li>
233		<a class="class" href="#Variable">Variable</a>
	329	<a class="class" href="#Group">Group</a>
234	330	<ul class="memberlist">
235	331	<li>
236		<a class="function" href="#Variable.__init__">Variable</a>
237		</li>
238		<li>
239		<a class="function" href="#Variable.group">group</a>
240		</li>
241		<li>
242		<a class="function" href="#Variable.ncattrs">ncattrs</a>
243		</li>
244		<li>
245		<a class="function" href="#Variable.setncattr">setncattr</a>
246		</li>
247		<li>
248		<a class="function" href="#Variable.setncattr_string">setncattr_string</a>
249		</li>
250		<li>
251		<a class="function" href="#Variable.setncatts">setncatts</a>
252		</li>
253		<li>
254		<a class="function" href="#Variable.getncattr">getncattr</a>
255		</li>
256		<li>
257		<a class="function" href="#Variable.delncattr">delncattr</a>
258		</li>
259		<li>
260		<a class="function" href="#Variable.filters">filters</a>
261		</li>
262		<li>
263		<a class="function" href="#Variable.endian">endian</a>
264		</li>
265		<li>
266		<a class="function" href="#Variable.chunking">chunking</a>
267		</li>
268		<li>
269		<a class="function" href="#Variable.get_var_chunk_cache">get_var_chunk_cache</a>
270		</li>
271		<li>
272		<a class="function" href="#Variable.set_var_chunk_cache">set_var_chunk_cache</a>
273		</li>
274		<li>
275		<a class="function" href="#Variable.renameAttribute">renameAttribute</a>
276		</li>
277		<li>
278		<a class="function" href="#Variable.assignValue">assignValue</a>
279		</li>
280		<li>
281		<a class="function" href="#Variable.getValue">getValue</a>
282		</li>
283		<li>
284		<a class="function" href="#Variable.set_auto_chartostring">set_auto_chartostring</a>
285		</li>
286		<li>
287		<a class="function" href="#Variable.use_nc_get_vars">use_nc_get_vars</a>
288		</li>
289		<li>
290		<a class="function" href="#Variable.set_auto_maskandscale">set_auto_maskandscale</a>
291		</li>
292		<li>
293		<a class="function" href="#Variable.set_auto_scale">set_auto_scale</a>
294		</li>
295		<li>
296		<a class="function" href="#Variable.set_auto_mask">set_auto_mask</a>
297		</li>
298		<li>
299		<a class="function" href="#Variable.set_always_mask">set_always_mask</a>
300		</li>
301		<li>
302		<a class="function" href="#Variable.set_ncstring_attrs">set_ncstring_attrs</a>
303		</li>
304		<li>
305		<a class="function" href="#Variable.set_collective">set_collective</a>
306		</li>
307		<li>
308		<a class="function" href="#Variable.get_dims">get_dims</a>
309		</li>
310		<li>
311		<a class="variable" href="#Variable.name">name</a>
312		</li>
313		<li>
314		<a class="variable" href="#Variable.datatype">datatype</a>
315		</li>
316		<li>
317		<a class="variable" href="#Variable.shape">shape</a>
318		</li>
319		<li>
320		<a class="variable" href="#Variable.size">size</a>
321		</li>
322		<li>
323		<a class="variable" href="#Variable.dimensions">dimensions</a>
324		</li>
325		<li>
326		<a class="variable" href="#Variable.ndim">ndim</a>
327		</li>
328		<li>
329		<a class="variable" href="#Variable.dtype">dtype</a>
330		</li>
331		<li>
332		<a class="variable" href="#Variable.mask">mask</a>
333		</li>
334		<li>
335		<a class="variable" href="#Variable.scale">scale</a>
336		</li>
337		<li>
338		<a class="variable" href="#Variable.always_mask">always_mask</a>
339		</li>
340		<li>
341		<a class="variable" href="#Variable.chartostring">chartostring</a>
	332	<a class="function" href="#Group.__init__">Group</a>
	333	</li>
	334	<li>
	335	<a class="function" href="#Group.close">close</a>
	336	</li>
	337	</ul>
	338
	339	</li>
	340	<li>
	341	<a class="class" href="#MFDataset">MFDataset</a>
	342	<ul class="memberlist">
	343	<li>
	344	<a class="function" href="#MFDataset.__init__">MFDataset</a>
	345	</li>
	346	<li>
	347	<a class="function" href="#MFDataset.ncattrs">ncattrs</a>
	348	</li>
	349	<li>
	350	<a class="function" href="#MFDataset.close">close</a>
	351	</li>
	352	</ul>
	353
	354	</li>
	355	<li>
	356	<a class="class" href="#MFTime">MFTime</a>
	357	<ul class="memberlist">
	358	<li>
	359	<a class="function" href="#MFTime.__init__">MFTime</a>
342	360	</li>
343	361	</ul>
344	362

377	395
378	396	</li>
379	397	<li>
380		<a class="class" href="#EnumType">EnumType</a>
381		<ul class="memberlist">
382		<li>
383		<a class="function" href="#EnumType.__init__">EnumType</a>
384		</li>
385		<li>
386		<a class="variable" href="#EnumType.dtype">dtype</a>
387		</li>
388		<li>
389		<a class="variable" href="#EnumType.name">name</a>
390		</li>
391		<li>
392		<a class="variable" href="#EnumType.enum_dict">enum_dict</a>
393		</li>
394		</ul>
395
	398	<a class="function" href="#date2num">date2num</a>
396	399	</li>
397	400	<li>
398		<a class="function" href="#getlibversion">getlibversion</a>
	401	<a class="function" href="#num2date">num2date</a>
399	402	</li>
400	403	<li>
401		<a class="function" href="#get_chunk_cache">get_chunk_cache</a>
402		</li>
403		<li>
404		<a class="function" href="#set_chunk_cache">set_chunk_cache</a>
405		</li>
406		<li>
407		<a class="function" href="#stringtoarr">stringtoarr</a>
	404	<a class="function" href="#date2index">date2index</a>
408	405	</li>
409	406	<li>
410	407	<a class="function" href="#stringtochar">stringtochar</a>

413	410	<a class="function" href="#chartostring">chartostring</a>
414	411	</li>
415	412	<li>
416		<a class="class" href="#MFDataset">MFDataset</a>
417		<ul class="memberlist">
418		<li>
419		<a class="function" href="#MFDataset.__init__">MFDataset</a>
420		</li>
421		<li>
422		<a class="function" href="#MFDataset.ncattrs">ncattrs</a>
423		</li>
424		<li>
425		<a class="function" href="#MFDataset.close">close</a>
426		</li>
427		</ul>
428
	413	<a class="function" href="#stringtoarr">stringtoarr</a>
429	414	</li>
430	415	<li>
431		<a class="class" href="#MFTime">MFTime</a>
	416	<a class="function" href="#getlibversion">getlibversion</a>
	417	</li>
	418	<li>
	419	<a class="class" href="#EnumType">EnumType</a>
432	420	<ul class="memberlist">
433	421	<li>
434		<a class="function" href="#MFTime.__init__">MFTime</a>
	422	<a class="function" href="#EnumType.__init__">EnumType</a>
	423	</li>
	424	<li>
	425	<a class="variable" href="#EnumType.dtype">dtype</a>
	426	</li>
	427	<li>
	428	<a class="variable" href="#EnumType.name">name</a>
	429	</li>
	430	<li>
	431	<a class="variable" href="#EnumType.enum_dict">enum_dict</a>
435	432	</li>
436	433	</ul>
437	434
	435	</li>
	436	<li>
	437	<a class="function" href="#get_chunk_cache">get_chunk_cache</a>
	438	</li>
	439	<li>
	440	<a class="function" href="#set_chunk_cache">set_chunk_cache</a>
438	441	</li>
439	442	</ul>
440	443
441	444
442	445	<a class="attribution" title="pdoc: Python API documentation generator" href="https://pdoc.dev">
443		built with <img
444		alt="pdoc"
	446	built with <span class="visually-hidden">pdoc</span><img
	447	alt="pdoc logo"
445	448	src="data:image/svg+xml,%3Csvg%20xmlns%3D%22http%3A//www.w3.org/2000/svg%22%20role%3D%22img%22%20aria-label%3D%22pdoc%20logo%22%20width%3D%22300%22%20height%3D%22150%22%20viewBox%3D%22-1%200%2060%2030%22%3E%3Ctitle%3Epdoc%3C/title%3E%3Cpath%20d%3D%22M29.621%2021.293c-.011-.273-.214-.475-.511-.481a.5.5%200%200%200-.489.503l-.044%201.393c-.097.551-.695%201.215-1.566%201.704-.577.428-1.306.486-2.193.182-1.426-.617-2.467-1.654-3.304-2.487l-.173-.172a3.43%203.43%200%200%200-.365-.306.49.49%200%200%200-.286-.196c-1.718-1.06-4.931-1.47-7.353.191l-.219.15c-1.707%201.187-3.413%202.131-4.328%201.03-.02-.027-.49-.685-.141-1.763.233-.721.546-2.408.772-4.076.042-.09.067-.187.046-.288.166-1.347.277-2.625.241-3.351%201.378-1.008%202.271-2.586%202.271-4.362%200-.976-.272-1.935-.788-2.774-.057-.094-.122-.18-.184-.268.033-.167.052-.339.052-.516%200-1.477-1.202-2.679-2.679-2.679-.791%200-1.496.352-1.987.9a6.3%206.3%200%200%200-1.001.029c-.492-.564-1.207-.929-2.012-.929-1.477%200-2.679%201.202-2.679%202.679A2.65%202.65%200%200%200%20.97%206.554c-.383.747-.595%201.572-.595%202.41%200%202.311%201.507%204.29%203.635%205.107-.037.699-.147%202.27-.423%203.294l-.137.461c-.622%202.042-2.515%208.257%201.727%2010.643%201.614.908%203.06%201.248%204.317%201.248%202.665%200%204.492-1.524%205.322-2.401%201.476-1.559%202.886-1.854%206.491.82%201.877%201.393%203.514%201.753%204.861%201.068%202.223-1.713%202.811-3.867%203.399-6.374.077-.846.056-1.469.054-1.537zm-4.835%204.313c-.054.305-.156.586-.242.629-.034-.007-.131-.022-.307-.157-.145-.111-.314-.478-.456-.908.221.121.432.25.675.355.115.039.219.051.33.081zm-2.251-1.238c-.05.33-.158.648-.252.694-.022.001-.125-.018-.307-.157-.217-.166-.488-.906-.639-1.573.358.344.754.693%201.198%201.036zm-3.887-2.337c-.006-.116-.018-.231-.041-.342.635.145%201.189.368%201.599.625.097.231.166.481.174.642-.03.049-.055.101-.067.158-.046.013-.128.026-.298.004-.278-.037-.901-.57-1.367-1.087zm-1.127-.497c.116.306.176.625.12.71-.019.014-.117.045-.345.016-.206-.027-.604-.332-.986-.695.41-.051.816-.056%201.211-.031zm-4.535%201.535c.209.22.379.47.358.598-.006.041-.088.138-.351.234-.144.055-.539-.063-.979-.259a11.66%2011.66%200%200%200%20.972-.573zm.983-.664c.359-.237.738-.418%201.126-.554.25.237.479.548.457.694-.006.042-.087.138-.351.235-.174.064-.694-.105-1.232-.375zm-3.381%201.794c-.022.145-.061.29-.149.401-.133.166-.358.248-.69.251h-.002c-.133%200-.306-.26-.45-.621.417.091.854.07%201.291-.031zm-2.066-8.077a4.78%204.78%200%200%201-.775-.584c.172-.115.505-.254.88-.378l-.105.962zm-.331%202.302a10.32%2010.32%200%200%201-.828-.502c.202-.143.576-.328.984-.49l-.156.992zm-.45%202.157l-.701-.403c.214-.115.536-.249.891-.376a11.57%2011.57%200%200%201-.19.779zm-.181%201.716c.064.398.194.702.298.893-.194-.051-.435-.162-.736-.398.061-.119.224-.3.438-.495zM8.87%204.141c0%20.152-.123.276-.276.276s-.275-.124-.275-.276.123-.276.276-.276.275.124.275.276zm-.735-.389a1.15%201.15%200%200%200-.314.783%201.16%201.16%200%200%200%201.162%201.162c.457%200%20.842-.27%201.032-.653.026.117.042.238.042.362a1.68%201.68%200%200%201-1.679%201.679%201.68%201.68%200%200%201-1.679-1.679c0-.843.626-1.535%201.436-1.654zM5.059%205.406A1.68%201.68%200%200%201%203.38%207.085a1.68%201.68%200%200%201-1.679-1.679c0-.037.009-.072.011-.109.21.3.541.508.935.508a1.16%201.16%200%200%200%201.162-1.162%201.14%201.14%200%200%200-.474-.912c.015%200%20.03-.005.045-.005.926.001%201.679.754%201.679%201.68zM3.198%204.141c0%20.152-.123.276-.276.276s-.275-.124-.275-.276.123-.276.276-.276.275.124.275.276zM1.375%208.964c0-.52.103-1.035.288-1.52.466.394%201.06.64%201.717.64%201.144%200%202.116-.725%202.499-1.738.383%201.012%201.355%201.738%202.499%201.738.867%200%201.631-.421%202.121-1.062.307.605.478%201.267.478%201.942%200%202.486-2.153%204.51-4.801%204.51s-4.801-2.023-4.801-4.51zm24.342%2019.349c-.985.498-2.267.168-3.813-.979-3.073-2.281-5.453-3.199-7.813-.705-1.315%201.391-4.163%203.365-8.423.97-3.174-1.786-2.239-6.266-1.261-9.479l.146-.492c.276-1.02.395-2.457.444-3.268a6.11%206.11%200%200%200%201.18.115%206.01%206.01%200%200%200%202.536-.562l-.006.175c-.802.215-1.848.612-2.021%201.25-.079.295.021.601.274.837.219.203.415.364.598.501-.667.304-1.243.698-1.311%201.179-.02.144-.022.507.393.787.213.144.395.26.564.365-1.285.521-1.361.96-1.381%201.126-.018.142-.011.496.427.746l.854.489c-.473.389-.971.914-.999%201.429-.018.278.095.532.316.713.675.556%201.231.721%201.653.721.059%200%20.104-.014.158-.02.207.707.641%201.64%201.513%201.64h.013c.8-.008%201.236-.345%201.462-.626.173-.216.268-.457.325-.692.424.195.93.374%201.372.374.151%200%20.294-.021.423-.068.732-.27.944-.704.993-1.021.009-.061.003-.119.002-.179.266.086.538.147.789.147.15%200%20.294-.021.423-.069.542-.2.797-.489.914-.754.237.147.478.258.704.288.106.014.205.021.296.021.356%200%20.595-.101.767-.229.438.435%201.094.992%201.656%201.067.106.014.205.021.296.021a1.56%201.56%200%200%200%20.323-.035c.17.575.453%201.289.866%201.605.358.273.665.362.914.362a.99.99%200%200%200%20.421-.093%201.03%201.03%200%200%200%20.245-.164c.168.428.39.846.68%201.068.358.273.665.362.913.362a.99.99%200%200%200%20.421-.093c.317-.148.512-.448.639-.762.251.157.495.257.726.257.127%200%20.25-.024.37-.071.427-.17.706-.617.841-1.314.022-.015.047-.022.068-.038.067-.051.133-.104.196-.159-.443%201.486-1.107%202.761-2.086%203.257zM8.66%209.925a.5.5%200%201%200-1%200c0%20.653-.818%201.205-1.787%201.205s-1.787-.552-1.787-1.205a.5.5%200%201%200-1%200c0%201.216%201.25%202.205%202.787%202.205s2.787-.989%202.787-2.205zm4.4%2015.965l-.208.097c-2.661%201.258-4.708%201.436-6.086.527-1.542-1.017-1.88-3.19-1.844-4.198a.4.4%200%200%200-.385-.414c-.242-.029-.406.164-.414.385-.046%201.249.367%203.686%202.202%204.896.708.467%201.547.7%202.51.7%201.248%200%202.706-.392%204.362-1.174l.185-.086a.4.4%200%200%200%20.205-.527c-.089-.204-.326-.291-.527-.206zM9.547%202.292c.093.077.205.114.317.114a.5.5%200%200%200%20.318-.886L8.817.397a.5.5%200%200%200-.703.068.5.5%200%200%200%20.069.703l1.364%201.124zm-7.661-.065c.086%200%20.173-.022.253-.068l1.523-.893a.5.5%200%200%200-.506-.863l-1.523.892a.5.5%200%200%200-.179.685c.094.158.261.247.432.247z%22%20transform%3D%22matrix%28-1%200%200%201%2058%200%29%22%20fill%3D%22%233bb300%22/%3E%3Cpath%20d%3D%22M.3%2021.86V10.18q0-.46.02-.68.04-.22.18-.5.28-.54%201.34-.54%201.06%200%201.42.28.38.26.44.78.76-1.04%202.38-1.04%201.64%200%203.1%201.54%201.46%201.54%201.46%203.58%200%202.04-1.46%203.58-1.44%201.54-3.08%201.54-1.64%200-2.38-.92v4.04q0%20.46-.04.68-.02.22-.18.5-.14.3-.5.42-.36.12-.98.12-.62%200-1-.12-.36-.12-.52-.4-.14-.28-.18-.5-.02-.22-.02-.68zm3.96-9.42q-.46.54-.46%201.18%200%20.64.46%201.18.48.52%201.2.52.74%200%201.24-.52.52-.52.52-1.18%200-.66-.48-1.18-.48-.54-1.26-.54-.76%200-1.22.54zm14.741-8.36q.16-.3.54-.42.38-.12%201-.12.64%200%201.02.12.38.12.52.42.16.3.18.54.04.22.04.68v11.94q0%20.46-.04.7-.02.22-.18.5-.3.54-1.7.54-1.38%200-1.54-.98-.84.96-2.34.96-1.8%200-3.28-1.56-1.48-1.58-1.48-3.66%200-2.1%201.48-3.68%201.5-1.58%203.28-1.58%201.48%200%202.3%201v-4.2q0-.46.02-.68.04-.24.18-.52zm-3.24%2010.86q.52.54%201.26.54.74%200%201.22-.54.5-.54.5-1.18%200-.66-.48-1.22-.46-.56-1.26-.56-.8%200-1.28.56-.48.54-.48%201.2%200%20.66.52%201.2zm7.833-1.2q0-2.4%201.68-3.96%201.68-1.56%203.84-1.56%202.16%200%203.82%201.56%201.66%201.54%201.66%203.94%200%201.66-.86%202.96-.86%201.28-2.1%201.9-1.22.6-2.54.6-1.32%200-2.56-.64-1.24-.66-2.1-1.92-.84-1.28-.84-2.88zm4.18%201.44q.64.48%201.3.48.66%200%201.32-.5.66-.5.66-1.48%200-.98-.62-1.46-.62-.48-1.34-.48-.72%200-1.34.5-.62.5-.62%201.48%200%20.96.64%201.46zm11.412-1.44q0%20.84.56%201.32.56.46%201.18.46.64%200%201.18-.36.56-.38.9-.38.6%200%201.46%201.06.46.58.46%201.04%200%20.76-1.1%201.42-1.14.8-2.8.8-1.86%200-3.58-1.34-.82-.64-1.34-1.7-.52-1.08-.52-2.36%200-1.3.52-2.34.52-1.06%201.34-1.7%201.66-1.32%203.54-1.32.76%200%201.48.22.72.2%201.06.4l.32.2q.36.24.56.38.52.4.52.92%200%20.5-.42%201.14-.72%201.1-1.38%201.1-.38%200-1.08-.44-.36-.34-1.04-.34-.66%200-1.24.48-.58.48-.58%201.34z%22%20fill%3D%22green%22/%3E%3C/svg%3E"/>
446	449	</a>
447	450	</div>

449	452	<main class="pdoc">
450	453	<section>
451	454	<h1 class="modulename">
452		<a href="./index.html">netCDF4</a>._netCDF4 </h1>
453
454		<div class="docstring"><h2 id="version-156">Version 1.5.6</h2>
	455	netCDF4 </h1>
	456
	457	<div class="docstring"><h2 id="version-157">Version 1.5.7</h2>
455	458
456	459	<h1 id="introduction">Introduction</h1>
457	460

474	477	compound types containing enums, or vlens containing compound
475	478	types) are not supported.</p>
476	479
477		<h2 id="download">Download</h2>
	480	<h2 id="quick-install">Quick Install</h2>
478	481
479	482	<ul>
480		<li>Latest bleeding-edge code from the
	483	<li>the easiest way to get going is to install via <code>pip install netCDF4</code>.
	484	(or if you use the <a href="http://conda.io">conda</a> package manager <code>conda install -c conda_forge netCDF4</code>).</li>
	485	</ul>
	486
	487	<h2 id="developer-install">Developer Install</h2>
	488
	489	<ul>
	490	<li>Clone the
481	491	<a href="http://github.com/Unidata/netcdf4-python">github repository</a>.</li>
482		<li>Latest <a href="https://pypi.python.org/pypi/netCDF4">releases</a>
483		(source code and binary installers).</li>
484		</ul>
485
486		<h2 id="requires">Requires</h2>
487
488		<ul>
489		<li>Python 3.6 or later.</li>
490		<li><a href="http://numpy.scipy.org">numpy array module</a>, version 1.10.0 or later.</li>
491		<li><a href="http://cython.org">Cython</a>, version 0.21 or later.</li>
492		<li><a href="https://pypi.python.org/pypi/setuptools">setuptools</a>, version 18.0 or
493		later.</li>
494		<li>The HDF5 C library version 1.8.4-patch1 or higher
495		from <a href="ftp://ftp.hdfgroup.org/HDF5/current/src"></a>.
496		<strong><em>netCDF version 4.4.1 or higher is recommended if using HDF5 1.10.x -
497		otherwise resulting files may be unreadable by clients using earlier
498		versions of HDF5. For netCDF < 4.4.1, HDF5 version 1.8.x is recommended.</em></strong>
499		Be sure to build with <code>--enable-hl --enable-shared</code>.</li>
500		<li><a href="http://curl.haxx.se/libcurl">Libcurl</a>, if you want
501		<a href="http://opendap.org">OPeNDAP</a> support.</li>
502		<li><a href="http://www.hdfgroup.org/products/hdf4">HDF4</a>, if you want
503		to be able to read HDF4 "Scientific Dataset" (SD) files.</li>
504		<li>The netCDF-4 C library from the <a href="https://github.com/Unidata/netcdf-c/releases">github releases
505		page</a>.
506		Version 4.1.1 or higher is required (4.2 or higher recommended).
507		Be sure to build with <code>--enable-netcdf-4 --enable-shared</code>, and set
508		<code>CPPFLAGS="-I $HDF5_DIR/include"</code> and <code>LDFLAGS="-L $HDF5_DIR/lib"</code>,
509		where <code>$HDF5_DIR</code> is the directory where HDF5 was installed.
510		If you want <a href="http://opendap.org">OPeNDAP</a> support, add <code>--enable-dap</code>.
511		If you want HDF4 SD support, add <code>--enable-hdf4</code> and add
512		the location of the HDF4 headers and library to <code>$CPPFLAGS</code> and <code>$LDFLAGS</code>.</li>
513		<li>for MPI parallel IO support, an MPI-enabled versions of the netcdf library
514		is required, as is the <a href="http://mpi4py.scipy.org">mpi4py</a> python module.
	492	<li>Make sure the dependencies are satisfied (Python 3.6 or later,
	493	<a href="http://numpy.scipy.org">numpy</a>,
	494	<a href="http://cython.org">Cython</a>,
	495	<a href="https://github.com/Unidata/cftime">cftime</a>,
	496	<a href="https://pypi.python.org/pypi/setuptools">setuptools</a>,
	497	the <a href="https://www.hdfgroup.org/solutions/hdf5/">HDF5 C library</a>,
	498	and the <a href="https://www.unidata.ucar.edu/software/netcdf/">netCDF C library</a>).
	499	For MPI parallel IO support, an MPI-enabled versions of the netcdf library
	500	is required, as is <a href="http://mpi4py.scipy.org">mpi4py</a>.
515	501	Parallel IO further depends on the existence of MPI-enabled HDF5 or the
516	502	<a href="https://parallel-netcdf.github.io/">PnetCDF</a> library.</li>
517		<li><a href="https://github.com/Unidata/cftime">cftime</a> for
518		time and date handling utility functions.</li>
519		</ul>
520
521		<h2 id="install">Install</h2>
522
523		<ul>
524		<li>install the requisite python modules and C libraries (see above). It's
525		easiest if all the C libs are built as shared libraries.</li>
526		<li>By default, the utility <code>nc-config</code>, installed with netcdf 4.1.2 or higher,
	503	<li>By default, the utility <code>nc-config</code> (installed with netcdf-c)
527	504	will be run used to determine where all the dependencies live.</li>
528	505	<li>If <code>nc-config</code> is not in your default <code>PATH</code>, you can set the <code>NETCDF4_DIR</code>
529	506	environment variable and <code>setup.py</code> will look in <code>$NETCDF4_DIR/bin</code>.

535	512	As a last resort, the library and include paths can be set via environment variables.
536	513	If you go this route, set <code>USE_NCCONFIG</code> and <code>USE_SETUPCFG</code> to 0, and specify
537	514	<code>NETCDF4_LIBDIR</code>, <code>NETCDF4_INCDIR</code>, <code>HDF5_LIBDIR</code> and <code>HDF5_INCDIR</code>.
538		Similarly, environment variables
539		(all capitalized) can be used to set the include and library paths for
540		<code>hdf4</code>, <code>szip</code>, <code>jpeg</code>, <code>curl</code> and <code>zlib</code>. If the dependencies are not found
	515	If the dependencies are not found
541	516	in any of the paths specified by environment variables, then standard locations
542	517	(such as <code>/usr</code> and <code>/usr/local</code>) are searched.</li>
543	518	<li>run <code>python setup.py build</code>, then <code>python setup.py install</code> (as root if
544		necessary). <code>pip install</code> can be used to install pre-compiled binary wheels from
545		<a href="https://pypi.org/project/netCDF4">pypi</a>.</li>
	519	necessary).</li>
546	520	<li>run the tests in the 'test' directory by running <code>python run_all.py</code>.</li>
547	521	</ul>
548	522

568	542
569	543	<h2 id="creatingopeningclosing-a-netcdf-file">Creating/Opening/Closing a netCDF file</h2>
570	544
571		<p>To create a netCDF file from python, you simply call the <a href="#Dataset">Dataset</a>
	545	<p>To create a netCDF file from python, you simply call the <code><a href="#Dataset">Dataset</a></code>
572	546	constructor. This is also the method used to open an existing netCDF
573	547	file. If the file is open for write access (<code>mode='w', 'r+'</code> or <code>'a'</code>), you may
574	548	write any type of data including new dimensions, groups, variables and

592	566	keyword in the <code><a href="#Dataset">Dataset</a></code> constructor. The default format is
593	567	<code>NETCDF4</code>. To see how a given file is formatted, you can examine the
594	568	<code>data_model</code> attribute. Closing the netCDF file is
595		accomplished via the <a href="#<a href="#Dataset.close">Dataset.close</a>"><a href="#Dataset.close">Dataset.close</a></a> method of the <a href="#Dataset">Dataset</a>
	569	accomplished via the <code><a href="#Dataset.close">Dataset.close</a></code> method of the <code><a href="#Dataset">Dataset</a></code>
596	570	instance.</p>
597	571
598	572	<p>Here's an example:</p>

605	579	</code></pre></div>
606	580
607	581	<p>Remote <a href="http://opendap.org">OPeNDAP</a>-hosted datasets can be accessed for
608		reading over http if a URL is provided to the <a href="#Dataset">Dataset</a> constructor instead of a
	582	reading over http if a URL is provided to the <code><a href="#Dataset">Dataset</a></code> constructor instead of a
609	583	filename. However, this requires that the netCDF library be built with
610	584	OPenDAP support, via the <code>--enable-dap</code> configure option (added in
611	585	version 4.0.1).</p>

615	589	<p>netCDF version 4 added support for organizing data in hierarchical
616	590	groups, which are analogous to directories in a filesystem. Groups serve
617	591	as containers for variables, dimensions and attributes, as well as other
618		groups. A <a href="#Dataset">Dataset</a> creates a special group, called
	592	groups. A <code><a href="#Dataset">Dataset</a></code> creates a special group, called
619	593	the 'root group', which is similar to the root directory in a unix
620		filesystem. To create <a href="#Group">Group</a> instances, use the
621		<a href="#<a href="#Dataset.createGroup">Dataset.createGroup</a>"><a href="#Dataset.createGroup">Dataset.createGroup</a></a> method of a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a>
622		instance. <a href="#<a href="#Dataset.createGroup">Dataset.createGroup</a>"><a href="#Dataset.createGroup">Dataset.createGroup</a></a> takes a single argument, a
623		python string containing the name of the new group. The new <a href="#Group">Group</a>
	594	filesystem. To create <code><a href="#Group">Group</a></code> instances, use the
	595	<code><a href="#Dataset.createGroup">Dataset.createGroup</a></code> method of a <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code>
	596	instance. <code><a href="#Dataset.createGroup">Dataset.createGroup</a></code> takes a single argument, a
	597	python string containing the name of the new group. The new <code><a href="#Group">Group</a></code>
624	598	instances contained within the root group can be accessed by name using
625		the <code>groups</code> dictionary attribute of the <a href="#Dataset">Dataset</a> instance. Only
	599	the <code>groups</code> dictionary attribute of the <code><a href="#Dataset">Dataset</a></code> instance. Only
626	600	<code>NETCDF4</code> formatted files support Groups, if you try to create a Group
627	601	in a netCDF 3 file you will get an error message.</p>
628	602

642	616	<span class="o">>>></span>
643	617	</code></pre></div>
644	618
645		<p>Groups can exist within groups in a <a href="#Dataset">Dataset</a>, just as directories
646		exist within directories in a unix filesystem. Each <a href="#Group">Group</a> instance
	619	<p>Groups can exist within groups in a <code><a href="#Dataset">Dataset</a></code>, just as directories
	620	exist within directories in a unix filesystem. Each <code><a href="#Group">Group</a></code> instance
647	621	has a <code>groups</code> attribute dictionary containing all of the group
648		instances contained within that group. Each <a href="#Group">Group</a> instance also has a
	622	instances contained within that group. Each <code><a href="#Group">Group</a></code> instance also has a
649	623	<code>path</code> attribute that contains a simulated unix directory path to
650	624	that group. To simplify the creation of nested groups, you can
651		use a unix-like path as an argument to <a href="#<a href="#Dataset.createGroup">Dataset.createGroup</a>"><a href="#Dataset.createGroup">Dataset.createGroup</a></a>.</p>
	625	use a unix-like path as an argument to <code><a href="#Dataset.createGroup">Dataset.createGroup</a></code>.</p>
652	626
653	627	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="n">fcstgrp1</span> <span class="o">=</span> <span class="n">rootgrp</span><span class="o">.</span><span class="n">createGroup</span><span class="p">(</span><span class="s2">"/forecasts/model1"</span><span class="p">)</span>
654	628	<span class="o">>>></span> <span class="n">fcstgrp2</span> <span class="o">=</span> <span class="n">rootgrp</span><span class="o">.</span><span class="n">createGroup</span><span class="p">(</span><span class="s2">"/forecasts/model2"</span><span class="p">)</span>

660	634	returned.</p>
661	635
662	636	<p>Here's an example that shows how to navigate all the groups in a
663		<a href="#Dataset">Dataset</a>. The function <code>walktree</code> is a Python generator that is used
664		to walk the directory tree. Note that printing the <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a>
	637	<code><a href="#Dataset">Dataset</a></code>. The function <code>walktree</code> is a Python generator that is used
	638	to walk the directory tree. Note that printing the <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code>
665	639	object yields summary information about it's contents.</p>
666	640
667	641	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="k">def</span> <span class="nf">walktree</span><span class="p">(</span><span class="n">top</span><span class="p">):</span>

707	681	before any variables can be created the dimensions they use must be
708	682	created first. A special case, not often used in practice, is that of a
709	683	scalar variable, which has no dimensions. A dimension is created using
710		the <a href="#<a href="#Dataset.createDimension">Dataset.createDimension</a>"><a href="#Dataset.createDimension">Dataset.createDimension</a></a> method of a <a href="#Dataset">Dataset</a>
711		or <a href="#Group">Group</a> instance. A Python string is used to set the name of the
	684	the <code><a href="#Dataset.createDimension">Dataset.createDimension</a></code> method of a <code><a href="#Dataset">Dataset</a></code>
	685	or <code><a href="#Group">Group</a></code> instance. A Python string is used to set the name of the
712	686	dimension, and an integer value is used to set the size. To create an
713	687	unlimited dimension (a dimension that can be appended to), the size
714	688	value is set to <code>None</code> or 0. In this example, there both the <code>time</code> and

722	696	<span class="o">>>></span> <span class="n">lon</span> <span class="o">=</span> <span class="n">rootgrp</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s2">"lon"</span><span class="p">,</span> <span class="mi">144</span><span class="p">)</span>
723	697	</code></pre></div>
724	698
725		<p>All of the <a href="#Dimension">Dimension</a> instances are stored in a python dictionary.</p>
	699	<p>All of the <code><a href="#Dimension">Dimension</a></code> instances are stored in a python dictionary.</p>
726	700
727	701	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="n">rootgrp</span><span class="o">.</span><span class="n">dimensions</span><span class="p">)</span>
728	702	<span class="p">{</span><span class="s1">'level'</span><span class="p">:</span> <span class="o"><</span><span class="k">class</span> <span class="err">'</span><span class="nc">netCDF4</span><span class="o">.</span><span class="n">_netCDF4</span><span class="o">.</span><span class="n">Dimension</span><span class="s1">'> (unlimited): name = '</span><span class="n">level</span><span class="s1">', size = 0, '</span><span class="n">time</span><span class="s1">': <class '</span><span class="n">netCDF4</span><span class="o">.</span><span class="n">_netCDF4</span><span class="o">.</span><span class="n">Dimension</span><span class="s1">'> (unlimited): name = '</span><span class="n">time</span><span class="s1">', size = 0, '</span><span class="n">lat</span><span class="s1">': <class '</span><span class="n">netCDF4</span><span class="o">.</span><span class="n">_netCDF4</span><span class="o">.</span><span class="n">Dimension</span><span class="s1">'>: name = '</span><span class="n">lat</span><span class="s1">', size = 73, '</span><span class="n">lon</span><span class="s1">': <class '</span><span class="n">netCDF4</span><span class="o">.</span><span class="n">_netCDF4</span><span class="o">.</span><span class="n">Dimension</span><span class="s1">'>: name = '</span><span class="n">lon</span><span class="s1">', size = 144}</span>
729	703	</code></pre></div>
730	704
731		<p>Using the python <code>len</code> function with a <a href="#Dimension">Dimension</a> instance returns
	705	<p>Using the python <code>len</code> function with a <code><a href="#Dimension">Dimension</a></code> instance returns
732	706	current size of that dimension.
733		<a href="#<a href="#Dimension.isunlimited">Dimension.isunlimited</a>"><a href="#Dimension.isunlimited">Dimension.isunlimited</a></a> method of a <a href="#Dimension">Dimension</a> instance
	707	<code><a href="#Dimension.isunlimited">Dimension.isunlimited</a></code> method of a <code><a href="#Dimension">Dimension</a></code> instance
734	708	be used to determine if the dimensions is unlimited, or appendable.</p>
735	709
736	710	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">lon</span><span class="p">))</span>

741	715	<span class="kc">True</span>
742	716	</code></pre></div>
743	717
744		<p>Printing the <a href="#Dimension">Dimension</a> object
	718	<p>Printing the <code><a href="#Dimension">Dimension</a></code> object
745	719	provides useful summary info, including the name and length of the dimension,
746	720	and whether it is unlimited.</p>
747	721

753	727	<span class="o"><</span><span class="k">class</span> <span class="err">'</span><span class="nc">netCDF4</span><span class="o">.</span><span class="n">_netCDF4</span><span class="o">.</span><span class="n">Dimension</span><span class="s1">'>: name = '</span><span class="n">lon</span><span class="s1">', size = 144</span>
754	728	</code></pre></div>
755	729
756		<p><a href="#Dimension">Dimension</a> names can be changed using the
757		<a href="#Datatset.renameDimension">Datatset.renameDimension</a> method of a <a href="#Dataset">Dataset</a> or
758		<a href="#Group">Group</a> instance.</p>
	730	<p><code><a href="#Dimension">Dimension</a></code> names can be changed using the
	731	<code>Datatset.renameDimension</code> method of a <code><a href="#Dataset">Dataset</a></code> or
	732	<code><a href="#Group">Group</a></code> instance.</p>
759	733
760	734	<h2 id="variables-in-a-netcdf-file">Variables in a netCDF file</h2>
761	735

763	737	supplied by the <a href="http://numpy.scipy.org">numpy module</a>. However,
764	738	unlike numpy arrays, netCDF4 variables can be appended to along one or
765	739	more 'unlimited' dimensions. To create a netCDF variable, use the
766		<a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a> method of a <a href="#Dataset">Dataset</a> or
767		<a href="#Group">Group</a> instance. The <a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a> method
	740	<code><a href="#Dataset.createVariable">Dataset.createVariable</a></code> method of a <code><a href="#Dataset">Dataset</a></code> or
	741	<code><a href="#Group">Group</a></code> instance. The <code><a href="#Dataset.createVariable">Dataset.createVariable</a></code>j method
768	742	has two mandatory arguments, the variable name (a Python string), and
769	743	the variable datatype. The variable's dimensions are given by a tuple
770	744	containing the dimension names (defined previously with
771		<a href="#<a href="#Dataset.createDimension">Dataset.createDimension</a>"><a href="#Dataset.createDimension">Dataset.createDimension</a></a>). To create a scalar
	745	<code><a href="#Dataset.createDimension">Dataset.createDimension</a></code>). To create a scalar
772	746	variable, simply leave out the dimensions keyword. The variable
773	747	primitive datatypes correspond to the dtype attribute of a numpy array.
774	748	You can specify the datatype as a numpy dtype object, or anything that

786	760	can only be used if the file format is <code>NETCDF4</code>.</p>
787	761
788	762	<p>The dimensions themselves are usually also defined as variables, called
789		coordinate variables. The <a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a>
790		method returns an instance of the <a href="#Variable">Variable</a> class whose methods can be
	763	coordinate variables. The <code><a href="#Dataset.createVariable">Dataset.createVariable</a></code>
	764	method returns an instance of the <code><a href="#Variable">Variable</a></code> class whose methods can be
791	765	used later to access and set variable data and attributes.</p>
792	766
793	767	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="n">times</span> <span class="o">=</span> <span class="n">rootgrp</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s2">"time"</span><span class="p">,</span><span class="s2">"f8"</span><span class="p">,(</span><span class="s2">"time"</span><span class="p">,))</span>

799	773	<span class="o">>>></span> <span class="n">temp</span><span class="o">.</span><span class="n">units</span> <span class="o">=</span> <span class="s2">"K"</span>
800	774	</code></pre></div>
801	775
802		<p>To get summary info on a <a href="#Variable">Variable</a> instance in an interactive session,
	776	<p>To get summary info on a <code><a href="#Variable">Variable</a></code> instance in an interactive session,
803	777	just print it.</p>
804	778
805	779	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="n">temp</span><span class="p">)</span>

818	792
819	793	<p>If the intermediate groups do not yet exist, they will be created.</p>
820	794
821		<p>You can also query a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> instance directly to obtain <a href="#Group">Group</a> or
822		<a href="#Variable">Variable</a> instances using paths.</p>
	795	<p>You can also query a <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> instance directly to obtain <code><a href="#Group">Group</a></code> or
	796	<code><a href="#Variable">Variable</a></code> instances using paths.</p>
823	797
824	798	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="n">rootgrp</span><span class="p">[</span><span class="s2">"/forecasts/model1"</span><span class="p">])</span> <span class="c1"># a Group instance</span>
825	799	<span class="o"><</span><span class="k">class</span> <span class="err">'</span><span class="nc">netCDF4</span><span class="o">.</span><span class="n">_netCDF4</span><span class="o">.</span><span class="n">Group</span><span class="s1">'></span>

836	810	<span class="n">filling</span> <span class="n">on</span><span class="p">,</span> <span class="n">default</span> <span class="n">_FillValue</span> <span class="n">of</span> <span class="mf">9.969209968386869e+36</span> <span class="n">used</span>
837	811	</code></pre></div>
838	812
839		<p>All of the variables in the <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> are stored in a
	813	<p>All of the variables in the <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> are stored in a
840	814	Python dictionary, in the same way as the dimensions:</p>
841	815
842	816	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="n">rootgrp</span><span class="o">.</span><span class="n">variables</span><span class="p">)</span>

864	838	<span class="n">filling</span> <span class="n">on</span><span class="p">,</span> <span class="n">default</span> <span class="n">_FillValue</span> <span class="n">of</span> <span class="mf">9.969209968386869e+36</span> <span class="n">used</span><span class="p">}</span>
865	839	</code></pre></div>
866	840
867		<p><a href="#Variable">Variable</a> names can be changed using the
868		<a href="#<a href="#Dataset.renameVariable">Dataset.renameVariable</a>"><a href="#Dataset.renameVariable">Dataset.renameVariable</a></a> method of a <a href="#Dataset">Dataset</a>
	841	<p><code><a href="#Variable">Variable</a></code> names can be changed using the
	842	<code><a href="#Dataset.renameVariable">Dataset.renameVariable</a></code> method of a <code><a href="#Dataset">Dataset</a></code>
869	843	instance.</p>
870	844
871	845	<p>Variables can be sliced similar to numpy arrays, but there are some differences. See

875	849
876	850	<p>There are two types of attributes in a netCDF file, global and variable.
877	851	Global attributes provide information about a group, or the entire
878		dataset, as a whole. <a href="#Variable">Variable</a> attributes provide information about
	852	dataset, as a whole. <code><a href="#Variable">Variable</a></code> attributes provide information about
879	853	one of the variables in a group. Global attributes are set by assigning
880		values to <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> instance variables. <a href="#Variable">Variable</a>
881		attributes are set by assigning values to <a href="#Variable">Variable</a> instances
	854	values to <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> instance variables. <code><a href="#Variable">Variable</a></code>
	855	attributes are set by assigning values to <code><a href="#Variable">Variable</a></code> instances
882	856	variables. Attributes can be strings, numbers or sequences. Returning to
883	857	our example,</p>
884	858

894	868	<span class="o">>>></span> <span class="n">times</span><span class="o">.</span><span class="n">calendar</span> <span class="o">=</span> <span class="s2">"gregorian"</span>
895	869	</code></pre></div>
896	870
897		<p>The <a href="#<a href="#Dataset.ncattrs">Dataset.ncattrs</a>"><a href="#Dataset.ncattrs">Dataset.ncattrs</a></a> method of a <a href="#Dataset">Dataset</a>, <a href="#Group">Group</a> or
898		<a href="#Variable">Variable</a> instance can be used to retrieve the names of all the netCDF
	871	<p>The <code><a href="#Dataset.ncattrs">Dataset.ncattrs</a></code> method of a <code><a href="#Dataset">Dataset</a></code>, <code><a href="#Group">Group</a></code> or
	872	<code><a href="#Variable">Variable</a></code> instance can be used to retrieve the names of all the netCDF
899	873	attributes. This method is provided as a convenience, since using the
900	874	built-in <code>dir</code> Python function will return a bunch of private methods
901	875	and attributes that cannot (or should not) be modified by the user.</p>

907	881	<span class="n">Global</span> <span class="n">attr</span> <span class="n">source</span> <span class="o">=</span> <span class="n">netCDF4</span> <span class="n">python</span> <span class="n">module</span> <span class="n">tutorial</span>
908	882	</code></pre></div>
909	883
910		<p>The <code>__dict__</code> attribute of a <a href="#Dataset">Dataset</a>, <a href="#Group">Group</a> or <a href="#Variable">Variable</a>
	884	<p>The <code>__dict__</code> attribute of a <code><a href="#Dataset">Dataset</a></code>, <code><a href="#Group">Group</a></code> or <code><a href="#Variable">Variable</a></code>
911	885	instance provides all the netCDF attribute name/value pairs in a python
912	886	dictionary:</p>
913	887

915	889	<span class="p">{</span><span class="s1">'description'</span><span class="p">:</span> <span class="s1">'bogus example script'</span><span class="p">,</span> <span class="s1">'history'</span><span class="p">:</span> <span class="s1">'Created Mon Jul 8 14:19:41 2019'</span><span class="p">,</span> <span class="s1">'source'</span><span class="p">:</span> <span class="s1">'netCDF4 python module tutorial'</span><span class="p">}</span>
916	890	</code></pre></div>
917	891
918		<p>Attributes can be deleted from a netCDF <a href="#Dataset">Dataset</a>, <a href="#Group">Group</a> or
919		<a href="#Variable">Variable</a> using the python <code>del</code> statement (i.e. <code>del grp.foo</code>
	892	<p>Attributes can be deleted from a netCDF <code><a href="#Dataset">Dataset</a></code>, <code><a href="#Group">Group</a></code> or
	893	<code><a href="#Variable">Variable</a></code> using the python <code>del</code> statement (i.e. <code>del grp.foo</code>
920	894	removes the attribute <code>foo</code> the the group <code>grp</code>).</p>
921	895
922	896	<h2 id="writing-data-to-and-retrieving-data-from-a-netcdf-variable">Writing data to and retrieving data from a netCDF variable</h2>
923	897
924		<p>Now that you have a netCDF <a href="#Variable">Variable</a> instance, how do you put data
	898	<p>Now that you have a netCDF <code><a href="#Variable">Variable</a></code> instance, how do you put data
925	899	into it? You can just treat it like an array and assign data to a slice.</p>
926	900
927		<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="kn">import</span> <span class="nn">numpy</span>
928		<span class="o">>>></span> <span class="n">lats</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="o">-</span><span class="mi">90</span><span class="p">,</span><span class="mi">91</span><span class="p">,</span><span class="mf">2.5</span><span class="p">)</span>
929		<span class="o">>>></span> <span class="n">lons</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="o">-</span><span class="mi">180</span><span class="p">,</span><span class="mi">180</span><span class="p">,</span><span class="mf">2.5</span><span class="p">)</span>
	901	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
	902	<span class="o">>>></span> <span class="n">lats</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="o">-</span><span class="mi">90</span><span class="p">,</span><span class="mi">91</span><span class="p">,</span><span class="mf">2.5</span><span class="p">)</span>
	903	<span class="o">>>></span> <span class="n">lons</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="o">-</span><span class="mi">180</span><span class="p">,</span><span class="mi">180</span><span class="p">,</span><span class="mf">2.5</span><span class="p">)</span>
930	904	<span class="o">>>></span> <span class="n">latitudes</span><span class="p">[:]</span> <span class="o">=</span> <span class="n">lats</span>
931	905	<span class="o">>>></span> <span class="n">longitudes</span><span class="p">[:]</span> <span class="o">=</span> <span class="n">lons</span>
932	906	<span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="s2">"latitudes =</span><span class="se">\n</span><span class="si">{}</span><span class="s2">"</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">latitudes</span><span class="p">[:]))</span>

940	914	<span class="mf">90.</span> <span class="p">]</span>
941	915	</code></pre></div>
942	916
943		<p>Unlike NumPy's array objects, netCDF <a href="#Variable">Variable</a>
	917	<p>Unlike NumPy's array objects, netCDF <code><a href="#Variable">Variable</a></code>
944	918	objects with unlimited dimensions will grow along those dimensions if you
945	919	assign data outside the currently defined range of indices.</p>
946	920

1012	986
1013	987	<p>By default, netcdf4-python returns numpy masked arrays with values equal to the
1014	988	<code>missing_value</code> or <code>_FillValue</code> variable attributes masked. The
1015		<a href="#<a href="#Dataset.set_auto_mask">Dataset.set_auto_mask</a>"><a href="#Dataset.set_auto_mask">Dataset.set_auto_mask</a></a> <a href="#Dataset">Dataset</a> and <a href="#Variable">Variable</a> methods
	989	<code><a href="#Dataset.set_auto_mask">Dataset.set_auto_mask</a></code> <code><a href="#Dataset">Dataset</a></code> and <code><a href="#Variable">Variable</a></code> methods
1016	990	can be used to disable this feature so that
1017	991	numpy arrays are always returned, with the missing values included. Prior to
1018	992	version 1.4.0 the default behavior was to only return masked arrays when the
1019	993	requested slice contained missing values. This behavior can be recovered
1020		using the <a href="#<a href="#Dataset.set_always_mask">Dataset.set_always_mask</a>"><a href="#Dataset.set_always_mask">Dataset.set_always_mask</a></a> method. If a masked array is
	994	using the <code><a href="#Dataset.set_always_mask">Dataset.set_always_mask</a></code> method. If a masked array is
1021	995	written to a netCDF variable, the masked elements are filled with the
1022	996	value specified by the <code>missing_value</code> attribute. If the variable has
1023	997	no <code>missing_value</code>, the <code>_FillValue</code> is used instead.</p>

1050	1024	<span class="n">real_datetime</span><span class="p">(</span><span class="mi">2001</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="mi">0</span><span class="p">)]</span>
1051	1025	</code></pre></div>
1052	1026
1053		<p><code>num2date</code> converts numeric values of time in the specified <code>units</code>
1054		and <code>calendar</code> to datetime objects, and <code>date2num</code> does the reverse.
	1027	<p><code><a href="#num2date">num2date</a></code> converts numeric values of time in the specified <code>units</code>
	1028	and <code>calendar</code> to datetime objects, and <code><a href="#date2num">date2num</a></code> does the reverse.
1055	1029	All the calendars currently defined in the
1056	1030	<a href="http://cfconventions.org">CF metadata convention</a> are supported.
1057		A function called <code>date2index</code> is also provided which returns the indices
	1031	A function called <code><a href="#date2index">date2index</a></code> is also provided which returns the indices
1058	1032	of a netCDF time variable corresponding to a sequence of datetime instances.</p>
1059	1033
1060	1034	<h2 id="reading-data-from-a-multi-file-netcdf-dataset">Reading data from a multi-file netCDF dataset</h2>
1061	1035
1062	1036	<p>If you want to read data from a variable that spans multiple netCDF files,
1063		you can use the <a href="#MFDataset">MFDataset</a> class to read the data as if it were
	1037	you can use the <code><a href="#MFDataset">MFDataset</a></code> class to read the data as if it were
1064	1038	contained in a single file. Instead of using a single filename to create
1065		a <a href="#Dataset">Dataset</a> instance, create a <a href="#MFDataset">MFDataset</a> instance with either a list
	1039	a <code><a href="#Dataset">Dataset</a></code> instance, create a <code><a href="#MFDataset">MFDataset</a></code> instance with either a list
1066	1040	of filenames, or a string with a wildcard (which is then converted to
1067	1041	a sorted list of files using the python glob module).
1068	1042	Variables in the list of files that share the same unlimited

1077	1051	<span class="o">...</span> <span class="k">with</span> <span class="n">Dataset</span><span class="p">(</span><span class="s2">"mftest</span><span class="si">%s</span><span class="s2">.nc"</span> <span class="o">%</span> <span class="n">nf</span><span class="p">,</span> <span class="s2">"w"</span><span class="p">,</span> <span class="nb">format</span><span class="o">=</span><span class="s2">"NETCDF4_CLASSIC"</span><span class="p">)</span> <span class="k">as</span> <span class="n">f</span><span class="p">:</span>
1078	1052	<span class="o">...</span> <span class="n">_</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s2">"x"</span><span class="p">,</span><span class="kc">None</span><span class="p">)</span>
1079	1053	<span class="o">...</span> <span class="n">x</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s2">"x"</span><span class="p">,</span><span class="s2">"i"</span><span class="p">,(</span><span class="s2">"x"</span><span class="p">,))</span>
1080		<span class="o">...</span> <span class="n">x</span><span class="p">[</span><span class="mi">0</span><span class="p">:</span><span class="mi">10</span><span class="p">]</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">nf</span><span class="o"></span><span class="mi">10</span><span class="p">,</span><span class="mi">10</span><span class="o"></span><span class="p">(</span><span class="n">nf</span><span class="o">+</span><span class="mi">1</span><span class="p">))</span>
1081		</code></pre></div>
1082
1083		<p>Now read all the files back in at once with <a href="#MFDataset">MFDataset</a></p>
	1054	<span class="o">...</span> <span class="n">x</span><span class="p">[</span><span class="mi">0</span><span class="p">:</span><span class="mi">10</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">nf</span><span class="o"></span><span class="mi">10</span><span class="p">,</span><span class="mi">10</span><span class="o"></span><span class="p">(</span><span class="n">nf</span><span class="o">+</span><span class="mi">1</span><span class="p">))</span>
	1055	</code></pre></div>
	1056
	1057	<p>Now read all the files back in at once with <code><a href="#MFDataset">MFDataset</a></code></p>
1084	1058
1085	1059	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="kn">from</span> <span class="nn">netCDF4</span> <span class="kn">import</span> <span class="n">MFDataset</span>
1086	1060	<span class="o">>>></span> <span class="n">f</span> <span class="o">=</span> <span class="n">MFDataset</span><span class="p">(</span><span class="s2">"mftest*nc"</span><span class="p">)</span>

1092	1066	<span class="mi">96</span> <span class="mi">97</span> <span class="mi">98</span> <span class="mi">99</span><span class="p">]</span>
1093	1067	</code></pre></div>
1094	1068
1095		<p>Note that <a href="#MFDataset">MFDataset</a> can only be used to read, not write, multi-file
	1069	<p>Note that <code><a href="#MFDataset">MFDataset</a></code> can only be used to read, not write, multi-file
1096	1070	datasets.</p>
1097	1071
1098	1072	<h2 id="efficient-compression-of-netcdf-variables">Efficient compression of netCDF variables</h2>
1099	1073
1100		<p>Data stored in netCDF 4 <a href="#Variable">Variable</a> objects can be compressed and
	1074	<p>Data stored in netCDF 4 <code><a href="#Variable">Variable</a></code> objects can be compressed and
1101	1075	decompressed on the fly. The parameters for the compression are
1102	1076	determined by the <code>zlib</code>, <code>complevel</code> and <code>shuffle</code> keyword arguments
1103		to the <a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a> method. To turn on
	1077	to the <code><a href="#Dataset.createVariable">Dataset.createVariable</a></code> method. To turn on
1104	1078	compression, set <code>zlib=True</code>. The <code>complevel</code> keyword regulates the
1105	1079	speed and efficiency of the compression (1 being fastest, but lowest
1106	1080	compression ratio, 9 being slowest but best compression ratio). The

1109	1083	compression by reordering the bytes. The shuffle filter can
1110	1084	significantly improve compression ratios, and is on by default. Setting
1111	1085	<code>fletcher32</code> keyword argument to
1112		<a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a> to <code>True</code> (it's <code>False</code> by
	1086	<code><a href="#Dataset.createVariable">Dataset.createVariable</a></code> to <code>True</code> (it's <code>False</code> by
1113	1087	default) enables the Fletcher32 checksum algorithm for error detection.
1114	1088	It's also possible to set the HDF5 chunking parameters and endian-ness
1115	1089	of the binary data stored in the HDF5 file with the <code>chunksizes</code>
1116	1090	and <code>endian</code> keyword arguments to
1117		<a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a>. These keyword arguments only
	1091	<code><a href="#Dataset.createVariable">Dataset.createVariable</a></code>. These keyword arguments only
1118	1092	are relevant for <code>NETCDF4</code> and <code>NETCDF4_CLASSIC</code> files (where the
1119	1093	underlying file format is HDF5) and are silently ignored if the file
1120	1094	format is <code>NETCDF3_CLASSIC</code>, <code>NETCDF3_64BIT_OFFSET</code> or <code>NETCDF3_64BIT_DATA</code>.</p>

1123	1097	example, it is temperature data that was measured with a precision of
1124	1098	0.1 degrees), you can dramatically improve zlib compression by
1125	1099	quantizing (or truncating) the data using the <code>least_significant_digit</code>
1126		keyword argument to <a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a>. The least
	1100	keyword argument to <code><a href="#Dataset.createVariable">Dataset.createVariable</a></code>. The least
1127	1101	significant digit is the power of ten of the smallest decimal place in
1128	1102	the data that is a reliable value. For example if the data has a
1129	1103	precision of 0.1, then setting <code>least_significant_digit=1</code> will cause

1135	1109
1136	1110	<p>In our example, try replacing the line</p>
1137	1111
1138		<pre><code>```python
1139		>>> temp = rootgrp.createVariable("temp","f4",("time","level","lat","lon",))
1140		</code></pre>
	1112	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="n">temp</span> <span class="o">=</span> <span class="n">rootgrp</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s2">"temp"</span><span class="p">,</span><span class="s2">"f4"</span><span class="p">,(</span><span class="s2">"time"</span><span class="p">,</span><span class="s2">"level"</span><span class="p">,</span><span class="s2">"lat"</span><span class="p">,</span><span class="s2">"lon"</span><span class="p">,))</span>
	1113	</code></pre></div>
1141	1114
1142	1115	<p>with</p>
1143	1116

1163	1136	information for a point by reading one variable, instead of reading
1164	1137	different parameters from different variables. Compound data types
1165	1138	are created from the corresponding numpy data type using the
1166		<a href="#<a href="#Dataset.createCompoundType">Dataset.createCompoundType</a>"><a href="#Dataset.createCompoundType">Dataset.createCompoundType</a></a> method of a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> instance.
	1139	<code><a href="#Dataset.createCompoundType">Dataset.createCompoundType</a></code> method of a <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> instance.
1167	1140	Since there is no native complex data type in netcdf, compound types are handy
1168	1141	for storing numpy complex arrays. Here's an example:</p>
1169	1142
1170	1143	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="n">f</span> <span class="o">=</span> <span class="n">Dataset</span><span class="p">(</span><span class="s2">"complex.nc"</span><span class="p">,</span><span class="s2">"w"</span><span class="p">)</span>
1171	1144	<span class="o">>>></span> <span class="n">size</span> <span class="o">=</span> <span class="mi">3</span> <span class="c1"># length of 1-d complex array</span>
1172	1145	<span class="o">>>></span> <span class="c1"># create sample complex data.</span>
1173		<span class="o">>>></span> <span class="n">datac</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">exp</span><span class="p">(</span><span class="mi">1</span><span class="n">j</span><span class="o">*</span><span class="p">(</span><span class="mf">1.</span><span class="o">+</span><span class="n">numpy</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">numpy</span><span class="o">.</span><span class="n">pi</span><span class="p">,</span> <span class="n">size</span><span class="p">)))</span>
	1146	<span class="o">>>></span> <span class="n">datac</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">exp</span><span class="p">(</span><span class="mi">1</span><span class="n">j</span><span class="o">*</span><span class="p">(</span><span class="mf">1.</span><span class="o">+</span><span class="n">np</span><span class="o">.</span><span class="n">linspace</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">np</span><span class="o">.</span><span class="n">pi</span><span class="p">,</span> <span class="n">size</span><span class="p">)))</span>
1174	1147	<span class="o">>>></span> <span class="c1"># create complex128 compound data type.</span>
1175		<span class="o">>>></span> <span class="n">complex128</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">dtype</span><span class="p">([(</span><span class="s2">"real"</span><span class="p">,</span><span class="n">numpy</span><span class="o">.</span><span class="n">float_64</span><span class="p">),(</span><span class="s2">"imag"</span><span class="p">,</span><span class="n">numpy</span><span class="o">.</span><span class="n">float_64</span><span class="p">)])</span>
	1148	<span class="o">>>></span> <span class="n">complex128</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">dtype</span><span class="p">([(</span><span class="s2">"real"</span><span class="p">,</span><span class="n">np</span><span class="o">.</span><span class="n">float64</span><span class="p">),(</span><span class="s2">"imag"</span><span class="p">,</span><span class="n">np</span><span class="o">.</span><span class="n">float64</span><span class="p">)])</span>
1176	1149	<span class="o">>>></span> <span class="n">complex128_t</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">createCompoundType</span><span class="p">(</span><span class="n">complex128</span><span class="p">,</span><span class="s2">"complex128"</span><span class="p">)</span>
1177	1150	<span class="o">>>></span> <span class="c1"># create a variable with this data type, write some data to it.</span>
1178	1151	<span class="o">>>></span> <span class="n">x_dim</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s2">"x_dim"</span><span class="p">,</span><span class="kc">None</span><span class="p">)</span>
1179	1152	<span class="o">>>></span> <span class="n">v</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s2">"cmplx_var"</span><span class="p">,</span><span class="n">complex128_t</span><span class="p">,</span><span class="s2">"x_dim"</span><span class="p">)</span>
1180		<span class="o">>>></span> <span class="n">data</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="n">size</span><span class="p">,</span><span class="n">complex128</span><span class="p">)</span> <span class="c1"># numpy structured array</span>
	1153	<span class="o">>>></span> <span class="n">data</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="n">size</span><span class="p">,</span><span class="n">complex128</span><span class="p">)</span> <span class="c1"># numpy structured array</span>
1181	1154	<span class="o">>>></span> <span class="n">data</span><span class="p">[</span><span class="s2">"real"</span><span class="p">]</span> <span class="o">=</span> <span class="n">datac</span><span class="o">.</span><span class="n">real</span><span class="p">;</span> <span class="n">data</span><span class="p">[</span><span class="s2">"imag"</span><span class="p">]</span> <span class="o">=</span> <span class="n">datac</span><span class="o">.</span><span class="n">imag</span>
1182	1155	<span class="o">>>></span> <span class="n">v</span><span class="p">[:]</span> <span class="o">=</span> <span class="n">data</span> <span class="c1"># write numpy structured array to netcdf compound var</span>
1183	1156	<span class="o">>>></span> <span class="c1"># close and reopen the file, check the contents.</span>

1185	1158	<span class="o">>>></span> <span class="n">v</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">variables</span><span class="p">[</span><span class="s2">"cmplx_var"</span><span class="p">]</span>
1186	1159	<span class="o">>>></span> <span class="n">datain</span> <span class="o">=</span> <span class="n">v</span><span class="p">[:]</span> <span class="c1"># read in all the data into a numpy structured array</span>
1187	1160	<span class="o">>>></span> <span class="c1"># create an empty numpy complex array</span>
1188		<span class="o">>>></span> <span class="n">datac2</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="n">datain</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span><span class="n">numpy</span><span class="o">.</span><span class="n">complex128</span><span class="p">)</span>
	1161	<span class="o">>>></span> <span class="n">datac2</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="n">datain</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span><span class="n">np</span><span class="o">.</span><span class="n">complex128</span><span class="p">)</span>
1189	1162	<span class="o">>>></span> <span class="c1"># .. fill it with contents of structured array.</span>
1190	1163	<span class="o">>>></span> <span class="n">datac2</span><span class="o">.</span><span class="n">real</span> <span class="o">=</span> <span class="n">datain</span><span class="p">[</span><span class="s2">"real"</span><span class="p">];</span> <span class="n">datac2</span><span class="o">.</span><span class="n">imag</span> <span class="o">=</span> <span class="n">datain</span><span class="p">[</span><span class="s2">"imag"</span><span class="p">]</span>
1191	1164	<span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="s1">'</span><span class="si">{}</span><span class="s1">: </span><span class="si">{}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">datac</span><span class="o">.</span><span class="n">dtype</span><span class="p">,</span> <span class="n">datac</span><span class="p">))</span> <span class="c1"># original data</span>

1199	1172	ones first. All possible numpy structured arrays cannot be
1200	1173	represented as Compound variables - an error message will be
1201	1174	raise if you try to create one that is not supported.
1202		All of the compound types defined for a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> are stored
	1175	All of the compound types defined for a <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> are stored
1203	1176	in a Python dictionary, just like variables and dimensions. As always, printing
1204	1177	objects gives useful summary information in an interactive session:</p>
1205	1178

1225	1198
1226	1199	<p>NetCDF 4 has support for variable-length or "ragged" arrays. These are arrays
1227	1200	of variable length sequences having the same type. To create a variable-length
1228		data type, use the <a href="#<a href="#Dataset.createVLType">Dataset.createVLType</a>"><a href="#Dataset.createVLType">Dataset.createVLType</a></a> method
1229		method of a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> instance.</p>
	1201	data type, use the <code><a href="#Dataset.createVLType">Dataset.createVLType</a></code> method
	1202	method of a <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> instance.</p>
1230	1203
1231	1204	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="n">f</span> <span class="o">=</span> <span class="n">Dataset</span><span class="p">(</span><span class="s2">"tst_vlen.nc"</span><span class="p">,</span><span class="s2">"w"</span><span class="p">)</span>
1232		<span class="o">>>></span> <span class="n">vlen_t</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">createVLType</span><span class="p">(</span><span class="n">numpy</span><span class="o">.</span><span class="n">int32</span><span class="p">,</span> <span class="s2">"phony_vlen"</span><span class="p">)</span>
	1205	<span class="o">>>></span> <span class="n">vlen_t</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">createVLType</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">int32</span><span class="p">,</span> <span class="s2">"phony_vlen"</span><span class="p">)</span>
1233	1206	</code></pre></div>
1234	1207
1235	1208	<p>The numpy datatype of the variable-length sequences and the name of the

1253	1226
1254	1227	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="kn">import</span> <span class="nn">random</span>
1255	1228	<span class="o">>>></span> <span class="n">random</span><span class="o">.</span><span class="n">seed</span><span class="p">(</span><span class="mi">54321</span><span class="p">)</span>
1256		<span class="o">>>></span> <span class="n">data</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">y</span><span class="p">)</span><span class="o">*</span><span class="nb">len</span><span class="p">(</span><span class="n">x</span><span class="p">),</span><span class="nb">object</span><span class="p">)</span>
	1229	<span class="o">>>></span> <span class="n">data</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">y</span><span class="p">)</span><span class="o">*</span><span class="nb">len</span><span class="p">(</span><span class="n">x</span><span class="p">),</span><span class="nb">object</span><span class="p">)</span>
1257	1230	<span class="o">>>></span> <span class="k">for</span> <span class="n">n</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">y</span><span class="p">)</span><span class="o">*</span><span class="nb">len</span><span class="p">(</span><span class="n">x</span><span class="p">)):</span>
1258		<span class="o">...</span> <span class="n">data</span><span class="p">[</span><span class="n">n</span><span class="p">]</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span><span class="mi">10</span><span class="p">),</span><span class="n">dtype</span><span class="o">=</span><span class="s2">"int32"</span><span class="p">)</span><span class="o">+</span><span class="mi">1</span>
1259		<span class="o">>>></span> <span class="n">data</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">data</span><span class="p">,(</span><span class="nb">len</span><span class="p">(</span><span class="n">y</span><span class="p">),</span><span class="nb">len</span><span class="p">(</span><span class="n">x</span><span class="p">)))</span>
	1231	<span class="o">...</span> <span class="n">data</span><span class="p">[</span><span class="n">n</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span><span class="mi">10</span><span class="p">),</span><span class="n">dtype</span><span class="o">=</span><span class="s2">"int32"</span><span class="p">)</span><span class="o">+</span><span class="mi">1</span>
	1232	<span class="o">>>></span> <span class="n">data</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="n">data</span><span class="p">,(</span><span class="nb">len</span><span class="p">(</span><span class="n">y</span><span class="p">),</span><span class="nb">len</span><span class="p">(</span><span class="n">x</span><span class="p">)))</span>
1260	1233	<span class="o">>>></span> <span class="n">vlvar</span><span class="p">[:]</span> <span class="o">=</span> <span class="n">data</span>
1261	1234	<span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="s2">"vlen variable =</span><span class="se">\n</span><span class="si">{}</span><span class="s2">"</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">vlvar</span><span class="p">[:]))</span>
1262	1235	<span class="n">vlen</span> <span class="n">variable</span> <span class="o">=</span>

1289	1262	variables, For vlen strings, you don't need to create a vlen data type.
1290	1263	Instead, simply use the python <code>str</code> builtin (or a numpy string datatype
1291	1264	with fixed length greater than 1) when calling the
1292		<a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a> method.</p>
	1265	<code><a href="#Dataset.createVariable">Dataset.createVariable</a></code> method.</p>
1293	1266
1294	1267	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="n">z</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s2">"z"</span><span class="p">,</span><span class="mi">10</span><span class="p">)</span>
1295	1268	<span class="o">>>></span> <span class="n">strvar</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s2">"strvar"</span><span class="p">,</span> <span class="nb">str</span><span class="p">,</span> <span class="s2">"z"</span><span class="p">)</span>

1300	1273	array is assigned to the vlen string variable.</p>
1301	1274
1302	1275	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="n">chars</span> <span class="o">=</span> <span class="s2">"1234567890aabcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"</span>
1303		<span class="o">>>></span> <span class="n">data</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span><span class="s2">"O"</span><span class="p">)</span>
	1276	<span class="o">>>></span> <span class="n">data</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span><span class="s2">"O"</span><span class="p">)</span>
1304	1277	<span class="o">>>></span> <span class="k">for</span> <span class="n">n</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">10</span><span class="p">):</span>
1305	1278	<span class="o">...</span> <span class="n">stringlen</span> <span class="o">=</span> <span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span><span class="mi">12</span><span class="p">)</span>
1306	1279	<span class="o">...</span> <span class="n">data</span><span class="p">[</span><span class="n">n</span><span class="p">]</span> <span class="o">=</span> <span class="s2">""</span><span class="o">.</span><span class="n">join</span><span class="p">([</span><span class="n">random</span><span class="o">.</span><span class="n">choice</span><span class="p">(</span><span class="n">chars</span><span class="p">)</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">stringlen</span><span class="p">)])</span>

1336	1309	<p>Here's an example of using an Enum type to hold cloud type data.
1337	1310	The base integer data type and a python dictionary describing the allowed
1338	1311	values and their names are used to define an Enum data type using
1339		<a href="#<a href="#Dataset.createEnumType">Dataset.createEnumType</a>"><a href="#Dataset.createEnumType">Dataset.createEnumType</a></a>.</p>
	1312	<code><a href="#Dataset.createEnumType">Dataset.createEnumType</a></code>.</p>
1340	1313
1341	1314	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="n">nc</span> <span class="o">=</span> <span class="n">Dataset</span><span class="p">(</span><span class="s1">'clouds.nc'</span><span class="p">,</span><span class="s1">'w'</span><span class="p">)</span>
1342	1315	<span class="o">>>></span> <span class="c1"># python dict with allowed values and their names.</span>

1345	1318	<span class="o">...</span> <span class="s1">'Nimbostratus'</span><span class="p">:</span> <span class="mi">6</span><span class="p">,</span> <span class="s1">'Cumulus'</span><span class="p">:</span> <span class="mi">4</span><span class="p">,</span> <span class="s1">'Altostratus'</span><span class="p">:</span> <span class="mi">5</span><span class="p">,</span>
1346	1319	<span class="o">...</span> <span class="s1">'Cumulonimbus'</span><span class="p">:</span> <span class="mi">1</span><span class="p">,</span> <span class="s1">'Stratocumulus'</span><span class="p">:</span> <span class="mi">3</span><span class="p">}</span>
1347	1320	<span class="o">>>></span> <span class="c1"># create the Enum type called 'cloud_t'.</span>
1348		<span class="o">>>></span> <span class="n">cloud_type</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createEnumType</span><span class="p">(</span><span class="n">numpy</span><span class="o">.</span><span class="n">uint8</span><span class="p">,</span><span class="s1">'cloud_t'</span><span class="p">,</span><span class="n">enum_dict</span><span class="p">)</span>
	1321	<span class="o">>>></span> <span class="n">cloud_type</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createEnumType</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">,</span><span class="s1">'cloud_t'</span><span class="p">,</span><span class="n">enum_dict</span><span class="p">)</span>
1349	1322	<span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="n">cloud_type</span><span class="p">)</span>
1350	1323	<span class="o"><</span><span class="k">class</span> <span class="err">'</span><span class="nc">netCDF4</span><span class="o">.</span><span class="n">_netCDF4</span><span class="o">.</span><span class="n">EnumType</span><span class="s1">'>: name = '</span><span class="n">cloud_t</span><span class="s1">', numpy dtype = uint8, fields/values ={'</span><span class="n">Altocumulus</span><span class="s1">': 7, '</span><span class="n">Missing</span><span class="s1">': 255, '</span><span class="n">Stratus</span><span class="s1">': 2, '</span><span class="n">Clear</span><span class="s1">': 0, '</span><span class="n">Nimbostratus</span><span class="s1">': 6, '</span><span class="n">Cumulus</span><span class="s1">': 4, '</span><span class="n">Altostratus</span><span class="s1">': 5, '</span><span class="n">Cumulonimbus</span><span class="s1">': 1, '</span><span class="n">Stratocumulus</span><span class="s1">': 3}</span>
1351	1324	</code></pre></div>

1414	1387	written to a different variable index on each task</p>
1415	1388
1416	1389	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="n">d</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s1">'dim'</span><span class="p">,</span><span class="mi">4</span><span class="p">)</span>
1417		<span class="o">>>></span> <span class="n">v</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s1">'var'</span><span class="p">,</span> <span class="n">np</span><span class="o">.</span><span class="n">int</span><span class="p">,</span> <span class="s1">'dim'</span><span class="p">)</span>
	1390	<span class="o">>>></span> <span class="n">v</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s1">'var'</span><span class="p">,</span> <span class="n">np</span><span class="o">.</span><span class="n">int64</span><span class="p">,</span> <span class="s1">'dim'</span><span class="p">)</span>
1418	1391	<span class="o">>>></span> <span class="n">v</span><span class="p">[</span><span class="n">rank</span><span class="p">]</span> <span class="o">=</span> <span class="n">rank</span>
1419	1392	<span class="o">>>></span> <span class="n">nc</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
1420	1393

1435	1408	depend on or be affected by other processes. Collective IO is a way of doing
1436	1409	IO defined in the MPI-IO standard; unlike independent IO, all processes must
1437	1410	participate in doing IO. To toggle back and forth between
1438		the two types of IO, use the <a href="#<a href="#Variable.set_collective">Variable.set_collective</a>"><a href="#Variable.set_collective">Variable.set_collective</a></a>
1439		<a href="#Variable">Variable</a> method. All metadata
	1411	the two types of IO, use the <code><a href="#Variable.set_collective">Variable.set_collective</a></code>
	1412	<code><a href="#Variable">Variable</a></code> method. All metadata
1440	1413	operations (such as creation of groups, types, variables, dimensions, or attributes)
1441	1414	are collective. There are a couple of important limitations of parallel IO:</p>
1442	1415

1449	1422	If the write is done in independent mode, the operation will fail with a
1450	1423	a generic "HDF Error".</li>
1451	1424	<li>You can write compressed data in parallel only with netcdf-c >= 4.7.4
1452		and hdf5 >= 1.10.3 (although you can read in parallel with earlier versions).</li>
	1425	and hdf5 >= 1.10.3 (although you can read in parallel with earlier versions). To write
	1426	compressed data in parallel, the variable must be in 'collective IO mode'. This is done
	1427	automatically on variable creation if compression is turned on, but if you are appending
	1428	to a variable in an existing file, you must use <code><a href="#Variable.set_collective">Variable.set_collective</a>(True)</code> before attempting
	1429	to write to it.</li>
1453	1430	<li>You cannot use variable-length (VLEN) data types.</li>
1454	1431	</ul>
1455	1432

1466	1443	To perform the conversion to and from character arrays to fixed-width numpy string arrays, the
1467	1444	following convention is followed by the python interface.
1468	1445	If the <code>_Encoding</code> special attribute is set for a character array
1469		(dtype <code>S1</code>) variable, the <a href="#chartostring">chartostring</a> utility function is used to convert the array of
	1446	(dtype <code>S1</code>) variable, the <code><a href="#chartostring">chartostring</a></code> utility function is used to convert the array of
1470	1447	characters to an array of strings with one less dimension (the last dimension is
1471	1448	interpreted as the length of each string) when reading the data. The character
1472	1449	set (usually ascii) is specified by the <code>_Encoding</code> attribute. If <code>_Encoding</code>
1473	1450	is 'none' or 'bytes', then the character array is converted to a numpy
1474	1451	fixed-width byte string array (dtype <code>S#</code>), otherwise a numpy unicode (dtype
1475	1452	<code>U#</code>) array is created. When writing the data,
1476		<a href="#stringtochar">stringtochar</a> is used to convert the numpy string array to an array of
	1453	<code><a href="#stringtochar">stringtochar</a></code> is used to convert the numpy string array to an array of
1477	1454	characters with one more dimension. For example,</p>
1478	1455
1479	1456	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="kn">from</span> <span class="nn">netCDF4</span> <span class="kn">import</span> <span class="n">stringtochar</span>

1481	1458	<span class="o">>>></span> <span class="n">_</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s1">'nchars'</span><span class="p">,</span><span class="mi">3</span><span class="p">)</span>
1482	1459	<span class="o">>>></span> <span class="n">_</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s1">'nstrings'</span><span class="p">,</span><span class="kc">None</span><span class="p">)</span>
1483	1460	<span class="o">>>></span> <span class="n">v</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s1">'strings'</span><span class="p">,</span><span class="s1">'S1'</span><span class="p">,(</span><span class="s1">'nstrings'</span><span class="p">,</span><span class="s1">'nchars'</span><span class="p">))</span>
1484		<span class="o">>>></span> <span class="n">datain</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">array</span><span class="p">([</span><span class="s1">'foo'</span><span class="p">,</span><span class="s1">'bar'</span><span class="p">],</span><span class="n">dtype</span><span class="o">=</span><span class="s1">'S3'</span><span class="p">)</span>
	1461	<span class="o">>>></span> <span class="n">datain</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">([</span><span class="s1">'foo'</span><span class="p">,</span><span class="s1">'bar'</span><span class="p">],</span><span class="n">dtype</span><span class="o">=</span><span class="s1">'S3'</span><span class="p">)</span>
1485	1462	<span class="o">>>></span> <span class="n">v</span><span class="p">[:]</span> <span class="o">=</span> <span class="n">stringtochar</span><span class="p">(</span><span class="n">datain</span><span class="p">)</span> <span class="c1"># manual conversion to char array</span>
1486	1463	<span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="n">v</span><span class="p">[:])</span> <span class="c1"># data returned as char array</span>
1487	1464	<span class="p">[[</span><span class="sa">b</span><span class="s1">'f'</span> <span class="sa">b</span><span class="s1">'o'</span> <span class="sa">b</span><span class="s1">'o'</span><span class="p">]</span>

1495	1472
1496	1473	<p>Even if the <code>_Encoding</code> attribute is set, the automatic conversion of char
1497	1474	arrays to/from string arrays can be disabled with
1498		<a href="#<a href="#Variable.set_auto_chartostring">Variable.set_auto_chartostring</a>"><a href="#Variable.set_auto_chartostring">Variable.set_auto_chartostring</a></a>.</p>
	1475	<code><a href="#Variable.set_auto_chartostring">Variable.set_auto_chartostring</a></code>.</p>
1499	1476
1500	1477	<p>A similar situation is often encountered with numpy structured arrays with subdtypes
1501	1478	containing fixed-wdith byte strings (dtype=<code>S#</code>). Since there is no native fixed-length string

1510	1487	Here's an example:</p>
1511	1488
1512	1489	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="n">nc</span> <span class="o">=</span> <span class="n">Dataset</span><span class="p">(</span><span class="s1">'compoundstring_example.nc'</span><span class="p">,</span><span class="s1">'w'</span><span class="p">)</span>
1513		<span class="o">>>></span> <span class="n">dtype</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">dtype</span><span class="p">([(</span><span class="s1">'observation'</span><span class="p">,</span> <span class="s1">'f4'</span><span class="p">),</span>
	1490	<span class="o">>>></span> <span class="n">dtype</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">dtype</span><span class="p">([(</span><span class="s1">'observation'</span><span class="p">,</span> <span class="s1">'f4'</span><span class="p">),</span>
1514	1491	<span class="o">...</span> <span class="p">(</span><span class="s1">'station_name'</span><span class="p">,</span><span class="s1">'S10'</span><span class="p">)])</span>
1515	1492	<span class="o">>>></span> <span class="n">station_data_t</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createCompoundType</span><span class="p">(</span><span class="n">dtype</span><span class="p">,</span><span class="s1">'station_data'</span><span class="p">)</span>
1516	1493	<span class="o">>>></span> <span class="n">_</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s1">'station'</span><span class="p">,</span><span class="kc">None</span><span class="p">)</span>
1517	1494	<span class="o">>>></span> <span class="n">statdat</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s1">'station_obs'</span><span class="p">,</span> <span class="n">station_data_t</span><span class="p">,</span> <span class="p">(</span><span class="s1">'station'</span><span class="p">,))</span>
1518		<span class="o">>>></span> <span class="n">data</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span><span class="n">dtype</span><span class="p">)</span>
	1495	<span class="o">>>></span> <span class="n">data</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">empty</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span><span class="n">dtype</span><span class="p">)</span>
1519	1496	<span class="o">>>></span> <span class="n">data</span><span class="p">[</span><span class="s1">'observation'</span><span class="p">][:]</span> <span class="o">=</span> <span class="p">(</span><span class="mf">123.</span><span class="p">,</span><span class="mf">3.14</span><span class="p">)</span>
1520	1497	<span class="o">>>></span> <span class="n">data</span><span class="p">[</span><span class="s1">'station_name'</span><span class="p">][:]</span> <span class="o">=</span> <span class="p">(</span><span class="s1">'Boulder'</span><span class="p">,</span><span class="s1">'New York'</span><span class="p">)</span>
1521	1498	<span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="n">statdat</span><span class="o">.</span><span class="n">dtype</span><span class="p">)</span> <span class="c1"># strings actually stored as character arrays</span>

1558	1535	<span class="o">>>></span> <span class="c1"># and persist the file to disk when it is closed.</span>
1559	1536	<span class="o">>>></span> <span class="n">nc</span> <span class="o">=</span> <span class="n">Dataset</span><span class="p">(</span><span class="s1">'diskless_example.nc'</span><span class="p">,</span><span class="s1">'w'</span><span class="p">,</span><span class="n">diskless</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span><span class="n">persist</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
1560	1537	<span class="o">>>></span> <span class="n">d</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s1">'x'</span><span class="p">,</span><span class="kc">None</span><span class="p">)</span>
1561		<span class="o">>>></span> <span class="n">v</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s1">'v'</span><span class="p">,</span><span class="n">numpy</span><span class="o">.</span><span class="n">int32</span><span class="p">,</span><span class="s1">'x'</span><span class="p">)</span>
1562		<span class="o">>>></span> <span class="n">v</span><span class="p">[</span><span class="mi">0</span><span class="p">:</span><span class="mi">5</span><span class="p">]</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">5</span><span class="p">)</span>
	1538	<span class="o">>>></span> <span class="n">v</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s1">'v'</span><span class="p">,</span><span class="n">np</span><span class="o">.</span><span class="n">int32</span><span class="p">,</span><span class="s1">'x'</span><span class="p">)</span>
	1539	<span class="o">>>></span> <span class="n">v</span><span class="p">[</span><span class="mi">0</span><span class="p">:</span><span class="mi">5</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">5</span><span class="p">)</span>
1563	1540	<span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="n">nc</span><span class="p">)</span>
1564	1541	<span class="o"><</span><span class="k">class</span> <span class="err">'</span><span class="nc">netCDF4</span><span class="o">.</span><span class="n">_netCDF4</span><span class="o">.</span><span class="n">Dataset</span><span class="s1">'></span>
1565	1542	<span class="n">root</span> <span class="n">group</span> <span class="p">(</span><span class="n">NETCDF4</span> <span class="n">data</span> <span class="n">model</span><span class="p">,</span> <span class="n">file</span> <span class="nb">format</span> <span class="n">HDF5</span><span class="p">):</span>

1592	1569	<span class="o">>>></span> <span class="c1"># (ignored for NETCDF4/HDF5 files).</span>
1593	1570	<span class="o">>>></span> <span class="n">nc</span> <span class="o">=</span> <span class="n">Dataset</span><span class="p">(</span><span class="s1">'inmemory.nc'</span><span class="p">,</span> <span class="n">mode</span><span class="o">=</span><span class="s1">'w'</span><span class="p">,</span><span class="n">memory</span><span class="o">=</span><span class="mi">1028</span><span class="p">)</span>
1594	1571	<span class="o">>>></span> <span class="n">d</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s1">'x'</span><span class="p">,</span><span class="kc">None</span><span class="p">)</span>
1595		<span class="o">>>></span> <span class="n">v</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s1">'v'</span><span class="p">,</span><span class="n">numpy</span><span class="o">.</span><span class="n">int32</span><span class="p">,</span><span class="s1">'x'</span><span class="p">)</span>
1596		<span class="o">>>></span> <span class="n">v</span><span class="p">[</span><span class="mi">0</span><span class="p">:</span><span class="mi">5</span><span class="p">]</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">5</span><span class="p">)</span>
	1572	<span class="o">>>></span> <span class="n">v</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">createVariable</span><span class="p">(</span><span class="s1">'v'</span><span class="p">,</span><span class="n">np</span><span class="o">.</span><span class="n">int32</span><span class="p">,</span><span class="s1">'x'</span><span class="p">)</span>
	1573	<span class="o">>>></span> <span class="n">v</span><span class="p">[</span><span class="mi">0</span><span class="p">:</span><span class="mi">5</span><span class="p">]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">5</span><span class="p">)</span>
1597	1574	<span class="o">>>></span> <span class="n">nc_buf</span> <span class="o">=</span> <span class="n">nc</span><span class="o">.</span><span class="n">close</span><span class="p">()</span> <span class="c1"># close returns memoryview</span>
1598	1575	<span class="o">>>></span> <span class="nb">print</span><span class="p">(</span><span class="nb">type</span><span class="p">(</span><span class="n">nc_buf</span><span class="p">))</span>
1599	1576	<span class="o"><</span><span class="k">class</span> <span class="err">'</span><span class="nc">memoryview</span><span class="s1">'></span>

1616	1593	the parallel IO example, which is in <code>examples/mpi_example.py</code>.
1617	1594	Unit tests are in the <code>test</code> directory.</p>
1618	1595
1619		<p><strong>contact</strong>: Jeffrey Whitaker <a href="mailto:jeffrey.s.whitaker@noaa.gov">jeffrey.s.whitaker@noaa.gov</a></p>
	1596	<p><strong>contact</strong>: Jeffrey Whitaker <a href="mailto:jeffrey.s.whitaker@noaa.gov">jeffrey.s.whitaker@noaa.gov</a></p>
1620	1597
1621	1598	<p><strong>copyright</strong>: 2008 by Jeffrey Whitaker.</p>
1622	1599

1627	1604	<p>THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.</p>
1628	1605	</div>
1629	1606
1630
	1607	<details>
	1608	<summary>View Source</summary>
	1609	<div class="codehilite"><pre><span></span><span class="c1"># init for netCDF4. package</span>
	1610	<span class="c1"># Docstring comes from extension module _netCDF4.</span>
	1611	<span class="kn">from</span> <span class="nn">._netCDF4</span> <span class="kn">import</span> <span class="o">*</span>
	1612	<span class="c1"># Need explicit imports for names beginning with underscores</span>
	1613	<span class="kn">from</span> <span class="nn">._netCDF4</span> <span class="kn">import</span> <span class="vm">__doc__</span>
	1614	<span class="kn">from</span> <span class="nn">._netCDF4</span> <span class="kn">import</span> <span class="p">(</span><span class="n">__version__</span><span class="p">,</span> <span class="n">__netcdf4libversion__</span><span class="p">,</span> <span class="n">__hdf5libversion__</span><span class="p">,</span>
	1615	<span class="n">__has_rename_grp__</span><span class="p">,</span> <span class="n">__has_nc_inq_path__</span><span class="p">,</span>
	1616	<span class="n">__has_nc_inq_format_extended__</span><span class="p">,</span> <span class="n">__has_nc_open_mem__</span><span class="p">,</span>
	1617	<span class="n">__has_nc_create_mem__</span><span class="p">,</span> <span class="n">__has_cdf5_format__</span><span class="p">,</span>
	1618	<span class="n">__has_parallel4_support__</span><span class="p">,</span> <span class="n">__has_pnetcdf_support__</span><span class="p">)</span>
	1619	<span class="n">__all__</span> <span class="o">=</span>\
	1620	<span class="p">[</span><span class="s1">'Dataset'</span><span class="p">,</span><span class="s1">'Variable'</span><span class="p">,</span><span class="s1">'Dimension'</span><span class="p">,</span><span class="s1">'Group'</span><span class="p">,</span><span class="s1">'MFDataset'</span><span class="p">,</span><span class="s1">'MFTime'</span><span class="p">,</span><span class="s1">'CompoundType'</span><span class="p">,</span><span class="s1">'VLType'</span><span class="p">,</span><span class="s1">'date2num'</span><span class="p">,</span><span class="s1">'num2date'</span><span class="p">,</span><span class="s1">'date2index'</span><span class="p">,</span><span class="s1">'stringtochar'</span><span class="p">,</span><span class="s1">'chartostring'</span><span class="p">,</span><span class="s1">'stringtoarr'</span><span class="p">,</span><span class="s1">'getlibversion'</span><span class="p">,</span><span class="s1">'EnumType'</span><span class="p">,</span><span class="s1">'get_chunk_cache'</span><span class="p">,</span><span class="s1">'set_chunk_cache'</span><span class="p">]</span>
	1621	</pre></div>
	1622
	1623	</details>
	1624
1631	1625	</section>
1632	1626	<section id="Dataset">
1633	1627	<div class="attr class">

1639	1633	</div>
1640	1634
1641	1635
1642		<div class="docstring"><p>A netCDF <a href="#Dataset">Dataset</a> is a collection of dimensions, groups, variables and
	1636	<div class="docstring"><p>A netCDF <code><a href="#Dataset">Dataset</a></code> is a collection of dimensions, groups, variables and
1643	1637	attributes. Together they describe the meaning of data and relations among
1644		data fields stored in a netCDF file. See <a href="#<a href="#Dataset.__init__">Dataset.__init__</a>"><a href="#Dataset.__init__">Dataset.__init__</a></a> for more
	1638	data fields stored in a netCDF file. See <code><a href="#Dataset.__init__">Dataset.__init__</a></code> for more
1645	1639	details.</p>
1646	1640
1647	1641	<p>A list of attribute names corresponding to global netCDF attributes
1648		defined for the <a href="#Dataset">Dataset</a> can be obtained with the
1649		<a href="#<a href="#Dataset.ncattrs">Dataset.ncattrs</a>"><a href="#Dataset.ncattrs">Dataset.ncattrs</a></a> method.
	1642	defined for the <code><a href="#Dataset">Dataset</a></code> can be obtained with the
	1643	<code><a href="#Dataset.ncattrs">Dataset.ncattrs</a></code> method.
1650	1644	These attributes can be created by assigning to an attribute of the
1651		<a href="#Dataset">Dataset</a> instance. A dictionary containing all the netCDF attribute
	1645	<code><a href="#Dataset">Dataset</a></code> instance. A dictionary containing all the netCDF attribute
1652	1646	name/value pairs is provided by the <code>__dict__</code> attribute of a
1653		<a href="#Dataset">Dataset</a> instance.</p>
	1647	<code><a href="#Dataset">Dataset</a></code> instance.</p>
1654	1648
1655	1649	<p>The following class variables are read-only and should not be
1656	1650	modified by the user.</p>
1657	1651
1658		<p><strong><code>dimensions</code></strong>: The <code>dimensions</code> dictionary maps the names of
1659		dimensions defined for the <a href="#Group">Group</a> or <a href="#Dataset">Dataset</a> to instances of the
1660		<a href="#Dimension">Dimension</a> class.</p>
1661
1662		<p><strong><code>variables</code></strong>: The <code>variables</code> dictionary maps the names of variables
1663		defined for this <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> to instances of the
1664		<a href="#Variable">Variable</a> class.</p>
1665
1666		<p><strong><code>groups</code></strong>: The groups dictionary maps the names of groups created for
1667		this <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> to instances of the <a href="#Group">Group</a> class (the
1668		<a href="#Dataset">Dataset</a> class is simply a special case of the <a href="#Group">Group</a> class which
	1652	<p><strong><code><a href="#Dataset.dimensions">dimensions</a></code></strong>: The <code><a href="#Dataset.dimensions">dimensions</a></code> dictionary maps the names of
	1653	dimensions defined for the <code><a href="#Group">Group</a></code> or <code><a href="#Dataset">Dataset</a></code> to instances of the
	1654	<code><a href="#Dimension">Dimension</a></code> class.</p>
	1655
	1656	<p><strong><code><a href="#Dataset.variables">variables</a></code></strong>: The <code><a href="#Dataset.variables">variables</a></code> dictionary maps the names of variables
	1657	defined for this <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> to instances of the
	1658	<code><a href="#Variable">Variable</a></code> class.</p>
	1659
	1660	<p><strong><code><a href="#Dataset.groups">groups</a></code></strong>: The groups dictionary maps the names of groups created for
	1661	this <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> to instances of the <code><a href="#Group">Group</a></code> class (the
	1662	<code><a href="#Dataset">Dataset</a></code> class is simply a special case of the <code><a href="#Group">Group</a></code> class which
1669	1663	describes the root group in the netCDF4 file).</p>
1670	1664
1671		<p><strong><code>cmptypes</code></strong>: The <code>cmptypes</code> dictionary maps the names of
1672		compound types defined for the <a href="#Group">Group</a> or <a href="#Dataset">Dataset</a> to instances of the
1673		<a href="#CompoundType">CompoundType</a> class.</p>
1674
1675		<p><strong><code>vltypes</code></strong>: The <code>vltypes</code> dictionary maps the names of
1676		variable-length types defined for the <a href="#Group">Group</a> or <a href="#Dataset">Dataset</a> to instances
1677		of the <a href="#VLType">VLType</a> class.</p>
1678
1679		<p><strong><code>enumtypes</code></strong>: The <code>enumtypes</code> dictionary maps the names of
1680		Enum types defined for the <a href="#Group">Group</a> or <a href="#Dataset">Dataset</a> to instances
1681		of the <a href="#EnumType">EnumType</a> class.</p>
1682
1683		<p><strong><code>data_model</code></strong>: <code>data_model</code> describes the netCDF
	1665	<p><strong><code><a href="#Dataset.cmptypes">cmptypes</a></code></strong>: The <code><a href="#Dataset.cmptypes">cmptypes</a></code> dictionary maps the names of
	1666	compound types defined for the <code><a href="#Group">Group</a></code> or <code><a href="#Dataset">Dataset</a></code> to instances of the
	1667	<code><a href="#CompoundType">CompoundType</a></code> class.</p>
	1668
	1669	<p><strong><code><a href="#Dataset.vltypes">vltypes</a></code></strong>: The <code><a href="#Dataset.vltypes">vltypes</a></code> dictionary maps the names of
	1670	variable-length types defined for the <code><a href="#Group">Group</a></code> or <code><a href="#Dataset">Dataset</a></code> to instances
	1671	of the <code><a href="#VLType">VLType</a></code> class.</p>
	1672
	1673	<p><strong><code><a href="#Dataset.enumtypes">enumtypes</a></code></strong>: The <code><a href="#Dataset.enumtypes">enumtypes</a></code> dictionary maps the names of
	1674	Enum types defined for the <code><a href="#Group">Group</a></code> or <code><a href="#Dataset">Dataset</a></code> to instances
	1675	of the <code><a href="#EnumType">EnumType</a></code> class.</p>
	1676
	1677	<p><strong><code><a href="#Dataset.data_model">data_model</a></code></strong>: <code><a href="#Dataset.data_model">data_model</a></code> describes the netCDF
1684	1678	data model version, one of <code>NETCDF3_CLASSIC</code>, <code>NETCDF4</code>,
1685	1679	<code>NETCDF4_CLASSIC</code>, <code>NETCDF3_64BIT_OFFSET</code> or <code>NETCDF3_64BIT_DATA</code>.</p>
1686	1680
1687		<p><strong><code>file_format</code></strong>: same as <code>data_model</code>, retained for backwards compatibility.</p>
1688
1689		<p><strong><code>disk_format</code></strong>: <code>disk_format</code> describes the underlying
	1681	<p><strong><code><a href="#Dataset.file_format">file_format</a></code></strong>: same as <code><a href="#Dataset.data_model">data_model</a></code>, retained for backwards compatibility.</p>
	1682
	1683	<p><strong><code><a href="#Dataset.disk_format">disk_format</a></code></strong>: <code><a href="#Dataset.disk_format">disk_format</a></code> describes the underlying
1690	1684	file format, one of <code>NETCDF3</code>, <code>HDF5</code>, <code>HDF4</code>,
1691	1685	<code>PNETCDF</code>, <code>DAP2</code>, <code>DAP4</code> or <code>UNDEFINED</code>. Only available if using
1692	1686	netcdf C library version >= 4.3.1, otherwise will always return
1693	1687	<code>UNDEFINED</code>.</p>
1694	1688
1695		<p><strong><code>parent</code></strong>: <code>parent</code> is a reference to the parent
1696		<a href="#Group">Group</a> instance. <code>None</code> for the root group or <a href="#Dataset">Dataset</a>
	1689	<p><strong><code><a href="#Dataset.parent">parent</a></code></strong>: <code><a href="#Dataset.parent">parent</a></code> is a reference to the parent
	1690	<code><a href="#Group">Group</a></code> instance. <code>None</code> for the root group or <code><a href="#Dataset">Dataset</a></code>
1697	1691	instance.</p>
1698	1692
1699		<p><strong><code>path</code></strong>: <code>path</code> shows the location of the <a href="#Group">Group</a> in
1700		the <a href="#Dataset">Dataset</a> in a unix directory format (the names of groups in the
1701		hierarchy separated by backslashes). A <a href="#Dataset">Dataset</a> instance is the root
	1693	<p><strong><code><a href="#Dataset.path">path</a></code></strong>: <code><a href="#Dataset.path">path</a></code> shows the location of the <code><a href="#Group">Group</a></code> in
	1694	the <code><a href="#Dataset">Dataset</a></code> in a unix directory format (the names of groups in the
	1695	hierarchy separated by backslashes). A <code><a href="#Dataset">Dataset</a></code> instance is the root
1702	1696	group, so the path is simply <code>'/'</code>.</p>
1703	1697
1704		<p><strong><code>keepweakref</code></strong>: If <code>True</code>, child Dimension and Variables objects only keep weak
	1698	<p><strong><code><a href="#Dataset.keepweakref">keepweakref</a></code></strong>: If <code>True</code>, child Dimension and Variables objects only keep weak
1705	1699	references to the parent Dataset or Group.</p>
1706	1700
1707	1701	<p><strong><code>_ncstring_attrs__</code></strong>: If <code>True</code>, all text attributes will be written as variable-length

1713	1707	<div class="attr function"><a class="headerlink" href="#Dataset.__init__"># &nbsp</a>
1714	1708
1715	1709
1716		<span class="name">Dataset</span><span class="signature">(unknown)</span> </div>
	1710	<span class="name">Dataset</span><span class="signature">()</span>
	1711	</div>
1717	1712
1718	1713
1719	1714	<div class="docstring"><p><strong><code>__init__(self, filename, mode="r", clobber=True, diskless=False,
1720	1715	persist=False, keepweakref=False, memory=None, encoding=None,
1721	1716	parallel=False, comm=None, info=None, format='NETCDF4')</code></strong></p>
1722	1717
1723		<p><a href="#Dataset">Dataset</a> constructor.</p>
	1718	<p><code><a href="#Dataset">Dataset</a></code> constructor.</p>
1724	1719
1725	1720	<p><strong><code>filename</code></strong>: Name of netCDF file to hold dataset. Can also
1726	1721	be a python 3 pathlib instance or the URL of an OpenDAP dataset. When memory is
1727		set this is just used to set the <code>filepath()</code>.</p>
	1722	set this is just used to set the <code><a href="#Dataset.filepath">filepath()</a></code>.</p>
1728	1723
1729	1724	<p><strong><code>mode</code></strong>: access mode. <code>r</code> means read-only; no data can be
1730	1725	modified. <code>w</code> means write; a new file is created, an existing file with

1768	1763	<p><strong><code>persist</code></strong>: if <code>diskless=True</code>, persist file to disk when closed
1769	1764	(default <code>False</code>).</p>
1770	1765
1771		<p><strong><code>keepweakref</code></strong>: if <code>True</code>, child Dimension and Variable instances will keep weak
	1766	<p><strong><code><a href="#Dataset.keepweakref">keepweakref</a></code></strong>: if <code>True</code>, child Dimension and Variable instances will keep weak
1772	1767	references to the parent Dataset or Group object. Default is <code>False</code>, which
1773	1768	means strong references will be kept. Having Dimension and Variable instances
1774	1769	keep a strong reference to the parent Dataset instance, which in turn keeps a

1872	1867
1873	1868	<div class="docstring"><p><strong><code>sync(self)</code></strong></p>
1874	1869
1875		<p>Writes all buffered data in the <a href="#Dataset">Dataset</a> to the disk file.</p>
	1870	<p>Writes all buffered data in the <code><a href="#Dataset">Dataset</a></code> to the disk file.</p>
1876	1871	</div>
1877	1872
1878	1873

1888	1883
1889	1884	<div class="docstring"><p><strong><code>set_fill_on(self)</code></strong></p>
1890	1885
1891		<p>Sets the fill mode for a <a href="#Dataset">Dataset</a> open for writing to <code>on</code>.</p>
	1886	<p>Sets the fill mode for a <code><a href="#Dataset">Dataset</a></code> open for writing to <code>on</code>.</p>
1892	1887
1893	1888	<p>This causes data to be pre-filled with fill values. The fill values can be
1894	1889	controlled by the variable's <code>_Fill_Value</code> attribute, but is usually
1895	1890	sufficient to the use the netCDF default <code>_Fill_Value</code> (defined
1896	1891	separately for each variable type). The default behavior of the netCDF
1897		library corresponds to <code>set_fill_on</code>. Data which are equal to the
	1892	library corresponds to <code><a href="#Dataset.set_fill_on">set_fill_on</a></code>. Data which are equal to the
1898	1893	<code>_Fill_Value</code> indicate that the variable was created, but never written
1899	1894	to.</p>
1900	1895	</div>

1912	1907
1913	1908	<div class="docstring"><p><strong><code>set_fill_off(self)</code></strong></p>
1914	1909
1915		<p>Sets the fill mode for a <a href="#Dataset">Dataset</a> open for writing to <code>off</code>.</p>
	1910	<p>Sets the fill mode for a <code><a href="#Dataset">Dataset</a></code> open for writing to <code>off</code>.</p>
1916	1911
1917	1912	<p>This will prevent the data from being pre-filled with fill values, which
1918	1913	may result in some performance improvements. However, you must then make

1936	1931
1937	1932	<p><code>size</code> must be a positive integer or <code>None</code>, which stands for
1938	1933	"unlimited" (default is <code>None</code>). Specifying a size of 0 also
1939		results in an unlimited dimension. The return value is the <a href="#Dimension">Dimension</a>
	1934	results in an unlimited dimension. The return value is the <code><a href="#Dimension">Dimension</a></code>
1940	1935	class instance describing the new dimension. To determine the current
1941		maximum size of the dimension, use the <code>len</code> function on the <a href="#Dimension">Dimension</a>
	1936	maximum size of the dimension, use the <code>len</code> function on the <code><a href="#Dimension">Dimension</a></code>
1942	1937	instance. To determine if a dimension is 'unlimited', use the
1943		<a href="#<a href="#Dimension.isunlimited">Dimension.isunlimited</a>"><a href="#Dimension.isunlimited">Dimension.isunlimited</a></a> method of the <a href="#Dimension">Dimension</a> instance.</p>
	1938	<code><a href="#Dimension.isunlimited">Dimension.isunlimited</a></code> method of the <code><a href="#Dimension">Dimension</a></code> instance.</p>
1944	1939	</div>
1945	1940
1946	1941

1956	1951
1957	1952	<div class="docstring"><p><strong><code>renameDimension(self, oldname, newname)</code></strong></p>
1958	1953
1959		<p>rename a <a href="#Dimension">Dimension</a> named <code>oldname</code> to <code>newname</code>.</p>
	1954	<p>rename a <code><a href="#Dimension">Dimension</a></code> named <code>oldname</code> to <code>newname</code>.</p>
1960	1955	</div>
1961	1956
1962	1957

1980	1975	are homogeneous numeric data types), then the 'inner' compound types <strong>must</strong> be
1981	1976	created first.</p>
1982	1977
1983		<p>The return value is the <a href="#CompoundType">CompoundType</a> class instance describing the new
	1978	<p>The return value is the <code><a href="#CompoundType">CompoundType</a></code> class instance describing the new
1984	1979	datatype.</p>
1985	1980	</div>
1986	1981

2000	1995	<p>Creates a new VLEN data type named <code>datatype_name</code> from a numpy
2001	1996	dtype object <code>datatype</code>.</p>
2002	1997
2003		<p>The return value is the <a href="#VLType">VLType</a> class instance describing the new
	1998	<p>The return value is the <code><a href="#VLType">VLType</a></code> class instance describing the new
2004	1999	datatype.</p>
2005	2000	</div>
2006	2001

2021	2016	integer dtype object <code>datatype</code>, and a python dictionary
2022	2017	defining the enum fields and values.</p>
2023	2018
2024		<p>The return value is the <a href="#EnumType">EnumType</a> class instance describing the new
	2019	<p>The return value is the <code><a href="#EnumType">EnumType</a></code> class instance describing the new
2025	2020	datatype.</p>
2026	2021	</div>
2027	2022

2041	2036	endian='native', least_significant_digit=None, fill_value=None, chunk_cache=None)</code></strong></p>
2042	2037
2043	2038	<p>Creates a new variable with the given <code>varname</code>, <code>datatype</code>, and
2044		<code>dimensions</code>. If dimensions are not given, the variable is assumed to be
	2039	<code><a href="#Dataset.dimensions">dimensions</a></code>. If dimensions are not given, the variable is assumed to be
2045	2040	a scalar.</p>
2046	2041
2047	2042	<p>If <code>varname</code> is specified as a path, using forward slashes as in unix to

2056	2051	(NC_BYTE), 'u1' (NC_UBYTE), 'i2' or 'h' or 's' (NC_SHORT), 'u2'
2057	2052	(NC_USHORT), 'i4' or 'i' or 'l' (NC_INT), 'u4' (NC_UINT), 'i8' (NC_INT64),
2058	2053	'u8' (NC_UINT64), 'f4' or 'f' (NC_FLOAT), 'f8' or 'd' (NC_DOUBLE)</code>.
2059		<code>datatype</code> can also be a <a href="#CompoundType">CompoundType</a> instance
2060		(for a structured, or compound array), a <a href="#VLType">VLType</a> instance
	2054	<code>datatype</code> can also be a <code><a href="#CompoundType">CompoundType</a></code> instance
	2055	(for a structured, or compound array), a <code><a href="#VLType">VLType</a></code> instance
2061	2056	(for a variable-length array), or the python <code>str</code> builtin
2062	2057	(for a variable-length string array). Numpy string and unicode datatypes with
2063	2058	length greater than one are aliases for <code>str</code>.</p>

2065	2060	<p>Data from netCDF variables is presented to python as numpy arrays with
2066	2061	the corresponding data type.</p>
2067	2062
2068		<p><code>dimensions</code> must be a tuple containing dimension names (strings) that
2069		have been defined previously using <a href="#<a href="#Dataset.createDimension">Dataset.createDimension</a>"><a href="#Dataset.createDimension">Dataset.createDimension</a></a>. The default value
	2063	<p><code><a href="#Dataset.dimensions">dimensions</a></code> must be a tuple containing dimension names (strings) that
	2064	have been defined previously using <code><a href="#Dataset.createDimension">Dataset.createDimension</a></code>. The default value
2070	2065	is an empty tuple, which means the variable is a scalar.</p>
2071	2066
2072	2067	<p>If the optional keyword <code>zlib</code> is <code>True</code>, the data will be compressed in

2108	2103
2109	2104	<p>The optional keyword <code>fill_value</code> can be used to override the default
2110	2105	netCDF <code>_FillValue</code> (the value that the variable gets filled with before
2111		any data is written to it, defaults given in <a href="#default_fillvals">default_fillvals</a>).
	2106	any data is written to it, defaults given in the dict <code><a href="#default_fillvals">netCDF4.default_fillvals</a></code>).
2112	2107	If fill_value is set to <code>False</code>, then the variable is not pre-filled.</p>
2113	2108
2114	2109	<p>If the optional keyword parameter <code>least_significant_digit</code> is

2133	2128	method to change it the next time the Dataset is opened.
2134	2129	Warning - messing with this parameter can seriously degrade performance.</p>
2135	2130
2136		<p>The return value is the <a href="#Variable">Variable</a> class instance describing the new
	2131	<p>The return value is the <code><a href="#Variable">Variable</a></code> class instance describing the new
2137	2132	variable.</p>
2138	2133
2139	2134	<p>A list of names corresponding to netCDF variable attributes can be
2140		obtained with the <a href="#Variable">Variable</a> method <a href="#<a href="#Variable.ncattrs">Variable.ncattrs</a>"><a href="#Variable.ncattrs">Variable.ncattrs</a></a>. A dictionary
	2135	obtained with the <code><a href="#Variable">Variable</a></code> method <code><a href="#Variable.ncattrs">Variable.ncattrs</a></code>. A dictionary
2141	2136	containing all the netCDF attribute name/value pairs is provided by
2142		the <code>__dict__</code> attribute of a <a href="#Variable">Variable</a> instance.</p>
2143
2144		<p><a href="#Variable">Variable</a> instances behave much like array objects. Data can be
	2137	the <code>__dict__</code> attribute of a <code><a href="#Variable">Variable</a></code> instance.</p>
	2138
	2139	<p><code><a href="#Variable">Variable</a></code> instances behave much like array objects. Data can be
2145	2140	assigned to or retrieved from a variable with indexing and slicing
2146		operations on the <a href="#Variable">Variable</a> instance. A <a href="#Variable">Variable</a> instance has six
	2141	operations on the <code><a href="#Variable">Variable</a></code> instance. A <code><a href="#Variable">Variable</a></code> instance has six
2147	2142	Dataset standard attributes: <code>dimensions, dtype, shape, ndim, name</code> and
2148	2143	<code>least_significant_digit</code>. Application programs should never modify
2149		these attributes. The <code>dimensions</code> attribute is a tuple containing the
	2144	these attributes. The <code><a href="#Dataset.dimensions">dimensions</a></code> attribute is a tuple containing the
2150	2145	names of the dimensions associated with this variable. The <code>dtype</code>
2151	2146	attribute is a string describing the variable's data type (<code>i4, f8,
2152	2147	S1,</code> etc). The <code>shape</code> attribute is a tuple describing the current
2153		sizes of all the variable's dimensions. The <code>name</code> attribute is a
	2148	sizes of all the variable's dimensions. The <code><a href="#Dataset.name">name</a></code> attribute is a
2154	2149	string containing the name of the Variable instance.
2155	2150	The <code>least_significant_digit</code>
2156	2151	attributes describes the power of ten of the smallest decimal place in
2157		the data the contains a reliable value. assigned to the <a href="#Variable">Variable</a>
	2152	the data the contains a reliable value. assigned to the <code><a href="#Variable">Variable</a></code>
2158	2153	instance. If <code>None</code>, the data is not truncated. The <code>ndim</code> attribute
2159	2154	is the number of variable dimensions.</p>
2160	2155	</div>

2172	2167
2173	2168	<div class="docstring"><p><strong><code>renameVariable(self, oldname, newname)</code></strong></p>
2174	2169
2175		<p>rename a <a href="#Variable">Variable</a> named <code>oldname</code> to <code>newname</code></p>
	2170	<p>rename a <code><a href="#Variable">Variable</a></code> named <code>oldname</code> to <code>newname</code></p>
2176	2171	</div>
2177	2172
2178	2173

2188	2183
2189	2184	<div class="docstring"><p><strong><code>createGroup(self, groupname)</code></strong></p>
2190	2185
2191		<p>Creates a new <a href="#Group">Group</a> with the given <code>groupname</code>.</p>
	2186	<p>Creates a new <code><a href="#Group">Group</a></code> with the given <code>groupname</code>.</p>
2192	2187
2193	2188	<p>If <code>groupname</code> is specified as a path, using forward slashes as in unix to
2194	2189	separate components, then intermediate groups will be created as necessary

2198	2193	If the specified path describes a group that already exists, no error is
2199	2194	raised.</p>
2200	2195
2201		<p>The return value is a <a href="#Group">Group</a> class instance.</p>
	2196	<p>The return value is a <code><a href="#Group">Group</a></code> class instance.</p>
2202	2197	</div>
2203	2198
2204	2199

2214	2209
2215	2210	<div class="docstring"><p><strong><code>ncattrs(self)</code></strong></p>
2216	2211
2217		<p>return netCDF global attribute names for this <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> in a list.</p>
	2212	<p>return netCDF global attribute names for this <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> in a list.</p>
2218	2213	</div>
2219	2214
2220	2215

2324	2319
2325	2320	<div class="docstring"><p><strong><code>renameAttribute(self, oldname, newname)</code></strong></p>
2326	2321
2327		<p>rename a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> attribute named <code>oldname</code> to <code>newname</code>.</p>
	2322	<p>rename a <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> attribute named <code>oldname</code> to <code>newname</code>.</p>
2328	2323	</div>
2329	2324
2330	2325

2340	2335
2341	2336	<div class="docstring"><p><strong><code>renameGroup(self, oldname, newname)</code></strong></p>
2342	2337
2343		<p>rename a <a href="#Group">Group</a> named <code>oldname</code> to <code>newname</code> (requires netcdf >= 4.3.1).</p>
	2338	<p>rename a <code><a href="#Group">Group</a></code> named <code>oldname</code> to <code>newname</code> (requires netcdf >= 4.3.1).</p>
2344	2339	</div>
2345	2340
2346	2341

2356	2351
2357	2352	<div class="docstring"><p><strong><code>set_auto_chartostring(self, True_or_False)</code></strong></p>
2358	2353
2359		<p>Call <a href="#<a href="#Variable.set_auto_chartostring">Variable.set_auto_chartostring</a>"><a href="#Variable.set_auto_chartostring">Variable.set_auto_chartostring</a></a> for all variables contained in this <a href="#Dataset">Dataset</a> or
2360		<a href="#Group">Group</a>, as well as for all variables in all its subgroups.</p>
	2354	<p>Call <code><a href="#Variable.set_auto_chartostring">Variable.set_auto_chartostring</a></code> for all variables contained in this <code><a href="#Dataset">Dataset</a></code> or
	2355	<code><a href="#Group">Group</a></code>, as well as for all variables in all its subgroups.</p>
2361	2356
2362	2357	<p><strong><code>True_or_False</code></strong>: Boolean determining if automatic conversion of
2363	2358	all character arrays <--> string arrays should be performed for

2381	2376
2382	2377	<div class="docstring"><p><strong><code>set_auto_maskandscale(self, True_or_False)</code></strong></p>
2383	2378
2384		<p>Call <a href="#<a href="#Variable.set_auto_maskandscale">Variable.set_auto_maskandscale</a>"><a href="#Variable.set_auto_maskandscale">Variable.set_auto_maskandscale</a></a> for all variables contained in this <a href="#Dataset">Dataset</a> or
2385		<a href="#Group">Group</a>, as well as for all variables in all its subgroups.</p>
	2379	<p>Call <code><a href="#Variable.set_auto_maskandscale">Variable.set_auto_maskandscale</a></code> for all variables contained in this <code><a href="#Dataset">Dataset</a></code> or
	2380	<code><a href="#Group">Group</a></code>, as well as for all variables in all its subgroups.</p>
2386	2381
2387	2382	<p><strong><code>True_or_False</code></strong>: Boolean determining if automatic conversion to masked arrays
2388	2383	and variable scaling shall be applied for all variables.</p>

2404	2399
2405	2400	<div class="docstring"><p><strong><code>set_auto_mask(self, True_or_False)</code></strong></p>
2406	2401
2407		<p>Call <a href="#<a href="#Variable.set_auto_mask">Variable.set_auto_mask</a>"><a href="#Variable.set_auto_mask">Variable.set_auto_mask</a></a> for all variables contained in this <a href="#Dataset">Dataset</a> or
2408		<a href="#Group">Group</a>, as well as for all variables in all its subgroups.</p>
	2402	<p>Call <code><a href="#Variable.set_auto_mask">Variable.set_auto_mask</a></code> for all variables contained in this <code><a href="#Dataset">Dataset</a></code> or
	2403	<code><a href="#Group">Group</a></code>, as well as for all variables in all its subgroups.</p>
2409	2404
2410	2405	<p><strong><code>True_or_False</code></strong>: Boolean determining if automatic conversion to masked arrays
2411	2406	shall be applied for all variables.</p>

2427	2422
2428	2423	<div class="docstring"><p><strong><code>set_auto_scale(self, True_or_False)</code></strong></p>
2429	2424
2430		<p>Call <a href="#<a href="#Variable.set_auto_scale">Variable.set_auto_scale</a>"><a href="#Variable.set_auto_scale">Variable.set_auto_scale</a></a> for all variables contained in this <a href="#Dataset">Dataset</a> or
2431		<a href="#Group">Group</a>, as well as for all variables in all its subgroups.</p>
	2425	<p>Call <code><a href="#Variable.set_auto_scale">Variable.set_auto_scale</a></code> for all variables contained in this <code><a href="#Dataset">Dataset</a></code> or
	2426	<code><a href="#Group">Group</a></code>, as well as for all variables in all its subgroups.</p>
2432	2427
2433	2428	<p><strong><code>True_or_False</code></strong>: Boolean determining if automatic variable scaling
2434	2429	shall be applied for all variables.</p>

2450	2445
2451	2446	<div class="docstring"><p><strong><code>set_always_mask(self, True_or_False)</code></strong></p>
2452	2447
2453		<p>Call <a href="#<a href="#Variable.set_always_mask">Variable.set_always_mask</a>"><a href="#Variable.set_always_mask">Variable.set_always_mask</a></a> for all variables contained in
2454		this <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a>, as well as for all
	2448	<p>Call <code><a href="#Variable.set_always_mask">Variable.set_always_mask</a></code> for all variables contained in
	2449	this <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code>, as well as for all
2455	2450	variables in all its subgroups.</p>
2456	2451
2457	2452	<p><strong><code>True_or_False</code></strong>: Boolean determining if automatic conversion of

2478	2473
2479	2474	<div class="docstring"><p><strong><code>set_ncstring_attrs(self, True_or_False)</code></strong></p>
2480	2475
2481		<p>Call <a href="#<a href="#Variable.set_ncstring_attrs">Variable.set_ncstring_attrs</a>"><a href="#Variable.set_ncstring_attrs">Variable.set_ncstring_attrs</a></a> for all variables contained in
2482		this <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a>, as well as for all its
	2476	<p>Call <code><a href="#Variable.set_ncstring_attrs">Variable.set_ncstring_attrs</a></code> for all variables contained in
	2477	this <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code>, as well as for all its
2483	2478	subgroups and their variables.</p>
2484	2479
2485	2480	<p><strong><code>True_or_False</code></strong>: Boolean determining if all string attributes are

2541	2536
2542	2537	<div class="docstring"><p><strong><code>fromcdl(cdlfilename, ncfilename=None, mode='a',format='NETCDF4')</code></strong></p>
2543	2538
2544		<p>call <code>ncgen</code> via subprocess to create Dataset from <a href="https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#cdl_guide">CDL</a>
2545		text representation. Requires <code>ncgen</code> to be installed and in <code>$PATH</code>.</p>
	2539	<p>call <a href="https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#ncgen_guide">ncgen</a> via subprocess to create Dataset from <a href="https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#cdl_guide">CDL</a>
	2540	text representation. Requires <a href="https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#ncgen_guide">ncgen</a> to be installed and in <code>$PATH</code>.</p>
2546	2541
2547	2542	<p><strong><code>cdlfilename</code></strong>: CDL file.</p>
2548	2543

2571	2566
2572	2567	<div class="docstring"><p><strong><code>tocdl(self, coordvars=False, data=False, outfile=None)</code></strong></p>
2573	2568
2574		<p>call <code>ncdump</code> via subprocess to create <a href="https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#cdl_guide">CDL</a>
2575		text representation of Dataset. Requires <code>ncdump</code> to be installed and in <code>$PATH</code>.</p>
	2569	<p>call <a href="https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#ncdump_guide">ncdump</a> via subprocess to create <a href="https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#cdl_guide">CDL</a>
	2570	text representation of Dataset. Requires <a href="https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#ncdump_guide">ncdump</a>
	2571	to be installed and in <code>$PATH</code>.</p>
2576	2572
2577	2573	<p><strong><code>coordvars</code></strong>: include coordinate variable data (via <code>ncdump -c</code>). Default False</p>
2578	2574

2586	2582	<div id="Dataset.name" class="classattr">
2587	2583	<div class="attr variable"><a class="headerlink" href="#Dataset.name"># &nbsp</a>
2588	2584
2589		<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2590		</div>
2591
2592
	2585	<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2586	</div>
	2587
	2588	<div class="docstring"><p>string name of Group instance</p>
	2589	</div>
	2590
2593	2591
2594	2592	</div>
2595	2593	<div id="Dataset.groups" class="classattr">
2596	2594	<div class="attr variable"><a class="headerlink" href="#Dataset.groups"># &nbsp</a>
2597	2595
2598		<span class="name">groups</span><span class="default_value"> = <attribute 'groups' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2596	<span class="name">groups</span><span class="default_value"> = <attribute 'groups' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2599	2597	</div>
2600	2598
2601	2599

2604	2602	<div id="Dataset.dimensions" class="classattr">
2605	2603	<div class="attr variable"><a class="headerlink" href="#Dataset.dimensions"># &nbsp</a>
2606	2604
2607		<span class="name">dimensions</span><span class="default_value"> = <attribute 'dimensions' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2605	<span class="name">dimensions</span><span class="default_value"> = <attribute 'dimensions' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2608	2606	</div>
2609	2607
2610	2608

2613	2611	<div id="Dataset.variables" class="classattr">
2614	2612	<div class="attr variable"><a class="headerlink" href="#Dataset.variables"># &nbsp</a>
2615	2613
2616		<span class="name">variables</span><span class="default_value"> = <attribute 'variables' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2614	<span class="name">variables</span><span class="default_value"> = <attribute 'variables' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2617	2615	</div>
2618	2616
2619	2617

2622	2620	<div id="Dataset.disk_format" class="classattr">
2623	2621	<div class="attr variable"><a class="headerlink" href="#Dataset.disk_format"># &nbsp</a>
2624	2622
2625		<span class="name">disk_format</span><span class="default_value"> = <attribute 'disk_format' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2623	<span class="name">disk_format</span><span class="default_value"> = <attribute 'disk_format' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2626	2624	</div>
2627	2625
2628	2626

2631	2629	<div id="Dataset.path" class="classattr">
2632	2630	<div class="attr variable"><a class="headerlink" href="#Dataset.path"># &nbsp</a>
2633	2631
2634		<span class="name">path</span><span class="default_value"> = <attribute 'path' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2632	<span class="name">path</span><span class="default_value"> = <attribute 'path' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2635	2633	</div>
2636	2634
2637	2635

2640	2638	<div id="Dataset.parent" class="classattr">
2641	2639	<div class="attr variable"><a class="headerlink" href="#Dataset.parent"># &nbsp</a>
2642	2640
2643		<span class="name">parent</span><span class="default_value"> = <attribute 'parent' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2641	<span class="name">parent</span><span class="default_value"> = <attribute 'parent' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2644	2642	</div>
2645	2643
2646	2644

2649	2647	<div id="Dataset.file_format" class="classattr">
2650	2648	<div class="attr variable"><a class="headerlink" href="#Dataset.file_format"># &nbsp</a>
2651	2649
2652		<span class="name">file_format</span><span class="default_value"> = <attribute 'file_format' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2650	<span class="name">file_format</span><span class="default_value"> = <attribute 'file_format' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2653	2651	</div>
2654	2652
2655	2653

2658	2656	<div id="Dataset.data_model" class="classattr">
2659	2657	<div class="attr variable"><a class="headerlink" href="#Dataset.data_model"># &nbsp</a>
2660	2658
2661		<span class="name">data_model</span><span class="default_value"> = <attribute 'data_model' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2659	<span class="name">data_model</span><span class="default_value"> = <attribute 'data_model' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2662	2660	</div>
2663	2661
2664	2662

2667	2665	<div id="Dataset.cmptypes" class="classattr">
2668	2666	<div class="attr variable"><a class="headerlink" href="#Dataset.cmptypes"># &nbsp</a>
2669	2667
2670		<span class="name">cmptypes</span><span class="default_value"> = <attribute 'cmptypes' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2668	<span class="name">cmptypes</span><span class="default_value"> = <attribute 'cmptypes' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2671	2669	</div>
2672	2670
2673	2671

2676	2674	<div id="Dataset.vltypes" class="classattr">
2677	2675	<div class="attr variable"><a class="headerlink" href="#Dataset.vltypes"># &nbsp</a>
2678	2676
2679		<span class="name">vltypes</span><span class="default_value"> = <attribute 'vltypes' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2677	<span class="name">vltypes</span><span class="default_value"> = <attribute 'vltypes' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2680	2678	</div>
2681	2679
2682	2680

2685	2683	<div id="Dataset.enumtypes" class="classattr">
2686	2684	<div class="attr variable"><a class="headerlink" href="#Dataset.enumtypes"># &nbsp</a>
2687	2685
2688		<span class="name">enumtypes</span><span class="default_value"> = <attribute 'enumtypes' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2686	<span class="name">enumtypes</span><span class="default_value"> = <attribute 'enumtypes' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2689	2687	</div>
2690	2688
2691	2689

2694	2692	<div id="Dataset.keepweakref" class="classattr">
2695	2693	<div class="attr variable"><a class="headerlink" href="#Dataset.keepweakref"># &nbsp</a>
2696	2694
2697		<span class="name">keepweakref</span><span class="default_value"> = <attribute 'keepweakref' of '<a href="#Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
	2695	<span class="name">keepweakref</span><span class="default_value"> = <attribute 'keepweakref' of '<a href="#_netCDF4.Dataset">netCDF4._netCDF4.Dataset</a>' objects></span>
2698	2696	</div>
2699	2697
2700	2698
	2699
	2700	</div>
	2701	</section>
	2702	<section id="Variable">
	2703	<div class="attr class">
	2704	<a class="headerlink" href="#Variable"># &nbsp</a>
	2705
	2706
	2707	<span class="def">class</span>
	2708	<span class="name">Variable</span>:
	2709	</div>
	2710
	2711
	2712	<div class="docstring"><p>A netCDF <code><a href="#Variable">Variable</a></code> is used to read and write netCDF data. They are
	2713	analogous to numpy array objects. See <code><a href="#Variable.__init__">Variable.__init__</a></code> for more
	2714	details.</p>
	2715
	2716	<p>A list of attribute names corresponding to netCDF attributes defined for
	2717	the variable can be obtained with the <code><a href="#Variable.ncattrs">Variable.ncattrs</a></code> method. These
	2718	attributes can be created by assigning to an attribute of the
	2719	<code><a href="#Variable">Variable</a></code> instance. A dictionary containing all the netCDF attribute
	2720	name/value pairs is provided by the <code>__dict__</code> attribute of a
	2721	<code><a href="#Variable">Variable</a></code> instance.</p>
	2722
	2723	<p>The following class variables are read-only:</p>
	2724
	2725	<p><strong><code><a href="#Variable.dimensions">dimensions</a></code></strong>: A tuple containing the names of the
	2726	dimensions associated with this variable.</p>
	2727
	2728	<p><strong><code><a href="#Variable.dtype">dtype</a></code></strong>: A numpy dtype object describing the
	2729	variable's data type.</p>
	2730
	2731	<p><strong><code><a href="#Variable.ndim">ndim</a></code></strong>: The number of variable dimensions.</p>
	2732
	2733	<p><strong><code><a href="#Variable.shape">shape</a></code></strong>: A tuple with the current shape (length of all dimensions).</p>
	2734
	2735	<p><strong><code><a href="#Variable.scale">scale</a></code></strong>: If True, <code>scale_factor</code> and <code>add_offset</code> are
	2736	applied, and signed integer data is automatically converted to
	2737	unsigned integer data if the <code>_Unsigned</code> attribute is set.
	2738	Default is <code>True</code>, can be reset using <code><a href="#Variable.set_auto_scale">Variable.set_auto_scale</a></code> and
	2739	<code><a href="#Variable.set_auto_maskandscale">Variable.set_auto_maskandscale</a></code> methods.</p>
	2740
	2741	<p><strong><code><a href="#Variable.mask">mask</a></code></strong>: If True, data is automatically converted to/from masked
	2742	arrays when missing values or fill values are present. Default is <code>True</code>, can be
	2743	reset using <code><a href="#Variable.set_auto_mask">Variable.set_auto_mask</a></code> and <code><a href="#Variable.set_auto_maskandscale">Variable.set_auto_maskandscale</a></code>
	2744	methods.</p>
	2745
	2746	<p><strong><code><a href="#Variable.chartostring">chartostring</a></code></strong>: If True, data is automatically converted to/from character
	2747	arrays to string arrays when the <code>_Encoding</code> variable attribute is set.
	2748	Default is <code>True</code>, can be reset using
	2749	<code><a href="#Variable.set_auto_chartostring">Variable.set_auto_chartostring</a></code> method.</p>
	2750
	2751	<p><strong><code>least_significant_digit</code></strong>: Describes the power of ten of the
	2752	smallest decimal place in the data the contains a reliable value. Data is
	2753	truncated to this decimal place when it is assigned to the <code><a href="#Variable">Variable</a></code>
	2754	instance. If <code>None</code>, the data is not truncated.</p>
	2755
	2756	<p><strong><code>__orthogonal_indexing__</code></strong>: Always <code>True</code>. Indicates to client code
	2757	that the object supports 'orthogonal indexing', which means that slices
	2758	that are 1d arrays or lists slice along each dimension independently. This
	2759	behavior is similar to Fortran or Matlab, but different than numpy.</p>
	2760
	2761	<p><strong><code><a href="#Variable.datatype">datatype</a></code></strong>: numpy data type (for primitive data types) or VLType/CompoundType
	2762	instance (for compound or vlen data types).</p>
	2763
	2764	<p><strong><code><a href="#Variable.name">name</a></code></strong>: String name.</p>
	2765
	2766	<p><strong><code><a href="#Variable.size">size</a></code></strong>: The number of stored elements.</p>
	2767	</div>
	2768
	2769
	2770	<div id="Variable.__init__" class="classattr">
	2771	<div class="attr function"><a class="headerlink" href="#Variable.__init__"># &nbsp</a>
	2772
	2773
	2774	<span class="name">Variable</span><span class="signature">()</span>
	2775	</div>
	2776
	2777
	2778	<div class="docstring"><p><strong><code>__init__(self, group, name, datatype, dimensions=(), zlib=False,
	2779	complevel=4, shuffle=True, fletcher32=False, contiguous=False,
	2780	chunksizes=None, endian='native',
	2781	least_significant_digit=None,fill_value=None,chunk_cache=None)</code></strong></p>
	2782
	2783	<p><code><a href="#Variable">Variable</a></code> constructor.</p>
	2784
	2785	<p><strong><code><a href="#Variable.group">group</a></code></strong>: <code><a href="#Group">Group</a></code> or <code><a href="#Dataset">Dataset</a></code> instance to associate with variable.</p>
	2786
	2787	<p><strong><code><a href="#Variable.name">name</a></code></strong>: Name of the variable.</p>
	2788
	2789	<p><strong><code><a href="#Variable.datatype">datatype</a></code></strong>: <code><a href="#Variable">Variable</a></code> data type. Can be specified by providing a
	2790	numpy dtype object, or a string that describes a numpy dtype object.
	2791	Supported values, corresponding to <code>str</code> attribute of numpy dtype
	2792	objects, include <code>'f4'</code> (32-bit floating point), <code>'f8'</code> (64-bit floating
	2793	point), <code>'i4'</code> (32-bit signed integer), <code>'i2'</code> (16-bit signed integer),
	2794	<code>'i8'</code> (64-bit signed integer), <code>'i4'</code> (8-bit signed integer), <code>'i1'</code>
	2795	(8-bit signed integer), <code>'u1'</code> (8-bit unsigned integer), <code>'u2'</code> (16-bit
	2796	unsigned integer), <code>'u4'</code> (32-bit unsigned integer), <code>'u8'</code> (64-bit
	2797	unsigned integer), or <code>'S1'</code> (single-character string). From
	2798	compatibility with Scientific.IO.NetCDF, the old Numeric single character
	2799	typecodes can also be used (<code>'f'</code> instead of <code>'f4'</code>, <code>'d'</code> instead of
	2800	<code>'f8'</code>, <code>'h'</code> or <code>'s'</code> instead of <code>'i2'</code>, <code>'b'</code> or <code>'B'</code> instead of
	2801	<code>'i1'</code>, <code>'c'</code> instead of <code>'S1'</code>, and <code>'i'</code> or <code>'l'</code> instead of
	2802	<code>'i4'</code>). <code><a href="#Variable.datatype">datatype</a></code> can also be a <code><a href="#CompoundType">CompoundType</a></code> instance
	2803	(for a structured, or compound array), a <code><a href="#VLType">VLType</a></code> instance
	2804	(for a variable-length array), or the python <code>str</code> builtin
	2805	(for a variable-length string array). Numpy string and unicode datatypes with
	2806	length greater than one are aliases for <code>str</code>.</p>
	2807
	2808	<p><strong><code><a href="#Variable.dimensions">dimensions</a></code></strong>: a tuple containing the variable's dimension names
	2809	(defined previously with <code>createDimension</code>). Default is an empty tuple
	2810	which means the variable is a scalar (and therefore has no dimensions).</p>
	2811
	2812	<p><strong><code>zlib</code></strong>: if <code>True</code>, data assigned to the <code><a href="#Variable">Variable</a></code>
	2813	instance is compressed on disk. Default <code>False</code>.</p>
	2814
	2815	<p><strong><code>complevel</code></strong>: the level of zlib compression to use (1 is the fastest,
	2816	but poorest compression, 9 is the slowest but best compression). Default 4.
	2817	Ignored if <code>zlib=False</code>.</p>
	2818
	2819	<p><strong><code>shuffle</code></strong>: if <code>True</code>, the HDF5 shuffle filter is applied
	2820	to improve compression. Default <code>True</code>. Ignored if <code>zlib=False</code>.</p>
	2821
	2822	<p><strong><code>fletcher32</code></strong>: if <code>True</code> (default <code>False</code>), the Fletcher32 checksum
	2823	algorithm is used for error detection.</p>
	2824
	2825	<p><strong><code>contiguous</code></strong>: if <code>True</code> (default <code>False</code>), the variable data is
	2826	stored contiguously on disk. Default <code>False</code>. Setting to <code>True</code> for
	2827	a variable with an unlimited dimension will trigger an error.</p>
	2828
	2829	<p><strong><code>chunksizes</code></strong>: Can be used to specify the HDF5 chunksizes for each
	2830	dimension of the variable. A detailed discussion of HDF chunking and I/O
	2831	performance is available
	2832	<a href="http://www.hdfgroup.org/HDF5/doc/H5.user/Chunking.html">here</a>.
	2833	Basically, you want the chunk size for each dimension to match as
	2834	closely as possible the size of the data block that users will read
	2835	from the file. <code>chunksizes</code> cannot be set if <code>contiguous=True</code>.</p>
	2836
	2837	<p><strong><code><a href="#Variable.endian">endian</a></code></strong>: Can be used to control whether the
	2838	data is stored in little or big endian format on disk. Possible
	2839	values are <code>little, big</code> or <code>native</code> (default). The library
	2840	will automatically handle endian conversions when the data is read,
	2841	but if the data is always going to be read on a computer with the
	2842	opposite format as the one used to create the file, there may be
	2843	some performance advantage to be gained by setting the endian-ness.
	2844	For netCDF 3 files (that don't use HDF5), only <code>endian='native'</code> is allowed.</p>
	2845
	2846	<p>The <code>zlib, complevel, shuffle, fletcher32, contiguous</code> and <code>chunksizes</code>
	2847	keywords are silently ignored for netCDF 3 files that do not use HDF5.</p>
	2848
	2849	<p><strong><code>least_significant_digit</code></strong>: If specified, variable data will be
	2850	truncated (quantized). In conjunction with <code>zlib=True</code> this produces
	2851	'lossy', but significantly more efficient compression. For example, if
	2852	<code>least_significant_digit=1</code>, data will be quantized using
	2853	around(scale<em>data)/scale, where scale = 2</em>*bits, and bits is determined
	2854	so that a precision of 0.1 is retained (in this case bits=4). Default is
	2855	<code>None</code>, or no quantization.</p>
	2856
	2857	<p><strong><code>fill_value</code></strong>: If specified, the default netCDF <code>_FillValue</code> (the
	2858	value that the variable gets filled with before any data is written to it)
	2859	is replaced with this value. If fill_value is set to <code>False</code>, then
	2860	the variable is not pre-filled. The default netCDF fill values can be found
	2861	in the dictionary <code><a href="#default_fillvals">netCDF4.default_fillvals</a></code>.</p>
	2862
	2863	<p><strong><code>chunk_cache</code></strong>: If specified, sets the chunk cache size for this variable.
	2864	Persists as long as Dataset is open. Use <code><a href="#Variable.set_var_chunk_cache">set_var_chunk_cache</a></code> to
	2865	change it when Dataset is re-opened. </p>
	2866
	2867	<p><strong><em>Note</em></strong>: <code><a href="#Variable">Variable</a></code> instances should be created using the
	2868	<code><a href="#Dataset.createVariable">Dataset.createVariable</a></code> method of a <code><a href="#Dataset">Dataset</a></code> or
	2869	<code><a href="#Group">Group</a></code> instance, not using this class directly.</p>
	2870	</div>
	2871
	2872
	2873	</div>
	2874	<div id="Variable.group" class="classattr">
	2875	<div class="attr function"><a class="headerlink" href="#Variable.group"># &nbsp</a>
	2876
	2877
	2878	<span class="def">def</span>
	2879	<span class="name">group</span><span class="signature">(unknown)</span>:
	2880	</div>
	2881
	2882
	2883	<div class="docstring"><p><strong><code>group(self)</code></strong></p>
	2884
	2885	<p>return the group that this <code><a href="#Variable">Variable</a></code> is a member of.</p>
	2886	</div>
	2887
	2888
	2889	</div>
	2890	<div id="Variable.ncattrs" class="classattr">
	2891	<div class="attr function"><a class="headerlink" href="#Variable.ncattrs"># &nbsp</a>
	2892
	2893
	2894	<span class="def">def</span>
	2895	<span class="name">ncattrs</span><span class="signature">(unknown)</span>:
	2896	</div>
	2897
	2898
	2899	<div class="docstring"><p><strong><code>ncattrs(self)</code></strong></p>
	2900
	2901	<p>return netCDF attribute names for this <code><a href="#Variable">Variable</a></code> in a list.</p>
	2902	</div>
	2903
	2904
	2905	</div>
	2906	<div id="Variable.setncattr" class="classattr">
	2907	<div class="attr function"><a class="headerlink" href="#Variable.setncattr"># &nbsp</a>
	2908
	2909
	2910	<span class="def">def</span>
	2911	<span class="name">setncattr</span><span class="signature">(unknown)</span>:
	2912	</div>
	2913
	2914
	2915	<div class="docstring"><p><strong><code>setncattr(self,name,value)</code></strong></p>
	2916
	2917	<p>set a netCDF variable attribute using name,value pair. Use if you need to set a
	2918	netCDF attribute with the same name as one of the reserved python
	2919	attributes.</p>
	2920	</div>
	2921
	2922
	2923	</div>
	2924	<div id="Variable.setncattr_string" class="classattr">
	2925	<div class="attr function"><a class="headerlink" href="#Variable.setncattr_string"># &nbsp</a>
	2926
	2927
	2928	<span class="def">def</span>
	2929	<span class="name">setncattr_string</span><span class="signature">(unknown)</span>:
	2930	</div>
	2931
	2932
	2933	<div class="docstring"><p><strong><code>setncattr_string(self,name,value)</code></strong></p>
	2934
	2935	<p>set a netCDF variable string attribute using name,value pair.
	2936	Use if you need to ensure that a netCDF attribute is created with type
	2937	<code>NC_STRING</code> if the file format is <code>NETCDF4</code>.
	2938	Use if you need to set an attribute to an array of variable-length strings.</p>
	2939	</div>
	2940
	2941
	2942	</div>
	2943	<div id="Variable.setncatts" class="classattr">
	2944	<div class="attr function"><a class="headerlink" href="#Variable.setncatts"># &nbsp</a>
	2945
	2946
	2947	<span class="def">def</span>
	2948	<span class="name">setncatts</span><span class="signature">(unknown)</span>:
	2949	</div>
	2950
	2951
	2952	<div class="docstring"><p><strong><code>setncatts(self,attdict)</code></strong></p>
	2953
	2954	<p>set a bunch of netCDF variable attributes at once using a python dictionary.
	2955	This may be faster when setting a lot of attributes for a <code>NETCDF3</code>
	2956	formatted file, since nc_redef/nc_enddef is not called in between setting
	2957	each attribute</p>
	2958	</div>
	2959
	2960
	2961	</div>
	2962	<div id="Variable.getncattr" class="classattr">
	2963	<div class="attr function"><a class="headerlink" href="#Variable.getncattr"># &nbsp</a>
	2964
	2965
	2966	<span class="def">def</span>
	2967	<span class="name">getncattr</span><span class="signature">(unknown)</span>:
	2968	</div>
	2969
	2970
	2971	<div class="docstring"><p><strong><code>getncattr(self,name)</code></strong></p>
	2972
	2973	<p>retrieve a netCDF variable attribute. Use if you need to set a
	2974	netCDF attribute with the same name as one of the reserved python
	2975	attributes.</p>
	2976
	2977	<p>option kwarg <code>encoding</code> can be used to specify the
	2978	character encoding of a string attribute (default is <code>utf-8</code>).</p>
	2979	</div>
	2980
	2981
	2982	</div>
	2983	<div id="Variable.delncattr" class="classattr">
	2984	<div class="attr function"><a class="headerlink" href="#Variable.delncattr"># &nbsp</a>
	2985
	2986
	2987	<span class="def">def</span>
	2988	<span class="name">delncattr</span><span class="signature">(unknown)</span>:
	2989	</div>
	2990
	2991
	2992	<div class="docstring"><p><strong><code>delncattr(self,name,value)</code></strong></p>
	2993
	2994	<p>delete a netCDF variable attribute. Use if you need to delete a
	2995	netCDF attribute with the same name as one of the reserved python
	2996	attributes.</p>
	2997	</div>
	2998
	2999
	3000	</div>
	3001	<div id="Variable.filters" class="classattr">
	3002	<div class="attr function"><a class="headerlink" href="#Variable.filters"># &nbsp</a>
	3003
	3004
	3005	<span class="def">def</span>
	3006	<span class="name">filters</span><span class="signature">(unknown)</span>:
	3007	</div>
	3008
	3009
	3010	<div class="docstring"><p><strong><code>filters(self)</code></strong></p>
	3011
	3012	<p>return dictionary containing HDF5 filter parameters.</p>
	3013	</div>
	3014
	3015
	3016	</div>
	3017	<div id="Variable.endian" class="classattr">
	3018	<div class="attr function"><a class="headerlink" href="#Variable.endian"># &nbsp</a>
	3019
	3020
	3021	<span class="def">def</span>
	3022	<span class="name">endian</span><span class="signature">(unknown)</span>:
	3023	</div>
	3024
	3025
	3026	<div class="docstring"><p><strong><code>endian(self)</code></strong></p>
	3027
	3028	<p>return endian-ness (<code>little,big,native</code>) of variable (as stored in HDF5 file).</p>
	3029	</div>
	3030
	3031
	3032	</div>
	3033	<div id="Variable.chunking" class="classattr">
	3034	<div class="attr function"><a class="headerlink" href="#Variable.chunking"># &nbsp</a>
	3035
	3036
	3037	<span class="def">def</span>
	3038	<span class="name">chunking</span><span class="signature">(unknown)</span>:
	3039	</div>
	3040
	3041
	3042	<div class="docstring"><p><strong><code>chunking(self)</code></strong></p>
	3043
	3044	<p>return variable chunking information. If the dataset is
	3045	defined to be contiguous (and hence there is no chunking) the word 'contiguous'
	3046	is returned. Otherwise, a sequence with the chunksize for
	3047	each dimension is returned.</p>
	3048	</div>
	3049
	3050
	3051	</div>
	3052	<div id="Variable.get_var_chunk_cache" class="classattr">
	3053	<div class="attr function"><a class="headerlink" href="#Variable.get_var_chunk_cache"># &nbsp</a>
	3054
	3055
	3056	<span class="def">def</span>
	3057	<span class="name">get_var_chunk_cache</span><span class="signature">(unknown)</span>:
	3058	</div>
	3059
	3060
	3061	<div class="docstring"><p><strong><code>get_var_chunk_cache(self)</code></strong></p>
	3062
	3063	<p>return variable chunk cache information in a tuple (size,nelems,preemption).
	3064	See netcdf C library documentation for <code>nc_get_var_chunk_cache</code> for
	3065	details.</p>
	3066	</div>
	3067
	3068
	3069	</div>
	3070	<div id="Variable.set_var_chunk_cache" class="classattr">
	3071	<div class="attr function"><a class="headerlink" href="#Variable.set_var_chunk_cache"># &nbsp</a>
	3072
	3073
	3074	<span class="def">def</span>
	3075	<span class="name">set_var_chunk_cache</span><span class="signature">(unknown)</span>:
	3076	</div>
	3077
	3078
	3079	<div class="docstring"><p><strong><code>set_var_chunk_cache(self,size=None,nelems=None,preemption=None)</code></strong></p>
	3080
	3081	<p>change variable chunk cache settings.
	3082	See netcdf C library documentation for <code>nc_set_var_chunk_cache</code> for
	3083	details.</p>
	3084	</div>
	3085
	3086
	3087	</div>
	3088	<div id="Variable.renameAttribute" class="classattr">
	3089	<div class="attr function"><a class="headerlink" href="#Variable.renameAttribute"># &nbsp</a>
	3090
	3091
	3092	<span class="def">def</span>
	3093	<span class="name">renameAttribute</span><span class="signature">(unknown)</span>:
	3094	</div>
	3095
	3096
	3097	<div class="docstring"><p><strong><code>renameAttribute(self, oldname, newname)</code></strong></p>
	3098
	3099	<p>rename a <code><a href="#Variable">Variable</a></code> attribute named <code>oldname</code> to <code>newname</code>.</p>
	3100	</div>
	3101
	3102
	3103	</div>
	3104	<div id="Variable.assignValue" class="classattr">
	3105	<div class="attr function"><a class="headerlink" href="#Variable.assignValue"># &nbsp</a>
	3106
	3107
	3108	<span class="def">def</span>
	3109	<span class="name">assignValue</span><span class="signature">(unknown)</span>:
	3110	</div>
	3111
	3112
	3113	<div class="docstring"><p><strong><code>assignValue(self, val)</code></strong></p>
	3114
	3115	<p>assign a value to a scalar variable. Provided for compatibility with
	3116	Scientific.IO.NetCDF, can also be done by assigning to an Ellipsis slice ([...]).</p>
	3117	</div>
	3118
	3119
	3120	</div>
	3121	<div id="Variable.getValue" class="classattr">
	3122	<div class="attr function"><a class="headerlink" href="#Variable.getValue"># &nbsp</a>
	3123
	3124
	3125	<span class="def">def</span>
	3126	<span class="name">getValue</span><span class="signature">(unknown)</span>:
	3127	</div>
	3128
	3129
	3130	<div class="docstring"><p><strong><code>getValue(self)</code></strong></p>
	3131
	3132	<p>get the value of a scalar variable. Provided for compatibility with
	3133	Scientific.IO.NetCDF, can also be done by slicing with an Ellipsis ([...]).</p>
	3134	</div>
	3135
	3136
	3137	</div>
	3138	<div id="Variable.set_auto_chartostring" class="classattr">
	3139	<div class="attr function"><a class="headerlink" href="#Variable.set_auto_chartostring"># &nbsp</a>
	3140
	3141
	3142	<span class="def">def</span>
	3143	<span class="name">set_auto_chartostring</span><span class="signature">(unknown)</span>:
	3144	</div>
	3145
	3146
	3147	<div class="docstring"><p><strong><code>set_auto_chartostring(self,chartostring)</code></strong></p>
	3148
	3149	<p>turn on or off automatic conversion of character variable data to and
	3150	from numpy fixed length string arrays when the <code>_Encoding</code> variable attribute
	3151	is set.</p>
	3152
	3153	<p>If <code><a href="#Variable.chartostring">chartostring</a></code> is set to <code>True</code>, when data is read from a character variable
	3154	(dtype = <code>S1</code>) that has an <code>_Encoding</code> attribute, it is converted to a numpy
	3155	fixed length unicode string array (dtype = <code>UN</code>, where <code>N</code> is the length
	3156	of the the rightmost dimension of the variable). The value of <code>_Encoding</code>
	3157	is the unicode encoding that is used to decode the bytes into strings.</p>
	3158
	3159	<p>When numpy string data is written to a variable it is converted back to
	3160	indiviual bytes, with the number of bytes in each string equalling the
	3161	rightmost dimension of the variable.</p>
	3162
	3163	<p>The default value of <code><a href="#Variable.chartostring">chartostring</a></code> is <code>True</code>
	3164	(automatic conversions are performed).</p>
	3165	</div>
	3166
	3167
	3168	</div>
	3169	<div id="Variable.use_nc_get_vars" class="classattr">
	3170	<div class="attr function"><a class="headerlink" href="#Variable.use_nc_get_vars"># &nbsp</a>
	3171
	3172
	3173	<span class="def">def</span>
	3174	<span class="name">use_nc_get_vars</span><span class="signature">(unknown)</span>:
	3175	</div>
	3176
	3177
	3178	<div class="docstring"><p><strong><code>use_nc_get_vars(self,_use_get_vars)</code></strong></p>
	3179
	3180	<p>enable the use of netcdf library routine <code>nc_get_vars</code>
	3181	to retrieve strided variable slices. By default,
	3182	<code>nc_get_vars</code> may not used by default (depending on the
	3183	version of the netcdf-c library being used) since it may be
	3184	slower than multiple calls to the unstrided read routine <code>nc_get_vara</code>.</p>
	3185	</div>
	3186
	3187
	3188	</div>
	3189	<div id="Variable.set_auto_maskandscale" class="classattr">
	3190	<div class="attr function"><a class="headerlink" href="#Variable.set_auto_maskandscale"># &nbsp</a>
	3191
	3192
	3193	<span class="def">def</span>
	3194	<span class="name">set_auto_maskandscale</span><span class="signature">(unknown)</span>:
	3195	</div>
	3196
	3197
	3198	<div class="docstring"><p><strong><code>set_auto_maskandscale(self,maskandscale)</code></strong></p>
	3199
	3200	<p>turn on or off automatic conversion of variable data to and
	3201	from masked arrays, automatic packing/unpacking of variable
	3202	data using <code>scale_factor</code> and <code>add_offset</code> attributes and
	3203	automatic conversion of signed integer data to unsigned integer
	3204	data if the <code>_Unsigned</code> attribute exists.</p>
	3205
	3206	<p>If <code>maskandscale</code> is set to <code>True</code>, when data is read from a variable
	3207	it is converted to a masked array if any of the values are exactly
	3208	equal to the either the netCDF _FillValue or the value specified by the
	3209	missing_value variable attribute. The fill_value of the masked array
	3210	is set to the missing_value attribute (if it exists), otherwise
	3211	the netCDF _FillValue attribute (which has a default value
	3212	for each data type). If the variable has no missing_value attribute, the
	3213	_FillValue is used instead. If the variable has valid_min/valid_max and
	3214	missing_value attributes, data outside the specified range will be masked.
	3215	When data is written to a variable, the masked
	3216	array is converted back to a regular numpy array by replacing all the
	3217	masked values by the missing_value attribute of the variable (if it
	3218	exists). If the variable has no missing_value attribute, the _FillValue
	3219	is used instead. </p>
	3220
	3221	<p>If <code>maskandscale</code> is set to <code>True</code>, and the variable has a
	3222	<code>scale_factor</code> or an <code>add_offset</code> attribute, then data read
	3223	from that variable is unpacked using::</p>
	3224
	3225	<pre><code>data = self.scale_factor*data + self.add_offset
	3226	</code></pre>
	3227
	3228	<p>When data is written to a variable it is packed using::</p>
	3229
	3230	<pre><code>data = (data - self.add_offset)/self.scale_factor
	3231	</code></pre>
	3232
	3233	<p>If either scale_factor is present, but add_offset is missing, add_offset
	3234	is assumed zero. If add_offset is present, but scale_factor is missing,
	3235	scale_factor is assumed to be one.
	3236	For more information on how <code>scale_factor</code> and <code>add_offset</code> can be
	3237	used to provide simple compression, see the
	3238	<a href="http://www.esrl.noaa.gov/psl/data/gridded/conventions/cdc_netcdf_standard.shtml">PSL metadata conventions</a>.</p>
	3239
	3240	<p>In addition, if <code>maskandscale</code> is set to <code>True</code>, and if the variable has an
	3241	attribute <code>_Unsigned</code> set, and the variable has a signed integer data type,
	3242	a view to the data is returned with the corresponding unsigned integer data type.
	3243	This convention is used by the netcdf-java library to save unsigned integer
	3244	data in <code>NETCDF3</code> or <code>NETCDF4_CLASSIC</code> files (since the <code>NETCDF3</code>
	3245	data model does not have unsigned integer data types).</p>
	3246
	3247	<p>The default value of <code>maskandscale</code> is <code>True</code>
	3248	(automatic conversions are performed).</p>
	3249	</div>
	3250
	3251
	3252	</div>
	3253	<div id="Variable.set_auto_scale" class="classattr">
	3254	<div class="attr function"><a class="headerlink" href="#Variable.set_auto_scale"># &nbsp</a>
	3255
	3256
	3257	<span class="def">def</span>
	3258	<span class="name">set_auto_scale</span><span class="signature">(unknown)</span>:
	3259	</div>
	3260
	3261
	3262	<div class="docstring"><p><strong><code>set_auto_scale(self,scale)</code></strong></p>
	3263
	3264	<p>turn on or off automatic packing/unpacking of variable
	3265	data using <code>scale_factor</code> and <code>add_offset</code> attributes.
	3266	Also turns on and off automatic conversion of signed integer data
	3267	to unsigned integer data if the variable has an <code>_Unsigned</code>
	3268	attribute.</p>
	3269
	3270	<p>If <code><a href="#Variable.scale">scale</a></code> is set to <code>True</code>, and the variable has a
	3271	<code>scale_factor</code> or an <code>add_offset</code> attribute, then data read
	3272	from that variable is unpacked using::</p>
	3273
	3274	<pre><code>data = self.scale_factor*data + self.add_offset
	3275	</code></pre>
	3276
	3277	<p>When data is written to a variable it is packed using::</p>
	3278
	3279	<pre><code>data = (data - self.add_offset)/self.scale_factor
	3280	</code></pre>
	3281
	3282	<p>If either scale_factor is present, but add_offset is missing, add_offset
	3283	is assumed zero. If add_offset is present, but scale_factor is missing,
	3284	scale_factor is assumed to be one.
	3285	For more information on how <code>scale_factor</code> and <code>add_offset</code> can be
	3286	used to provide simple compression, see the
	3287	<a href="http://www.esrl.noaa.gov/psl/data/gridded/conventions/cdc_netcdf_standard.shtml">PSL metadata conventions</a>.</p>
	3288
	3289	<p>In addition, if <code><a href="#Variable.scale">scale</a></code> is set to <code>True</code>, and if the variable has an
	3290	attribute <code>_Unsigned</code> set, and the variable has a signed integer data type,
	3291	a view to the data is returned with the corresponding unsigned integer datatype.
	3292	This convention is used by the netcdf-java library to save unsigned integer
	3293	data in <code>NETCDF3</code> or <code>NETCDF4_CLASSIC</code> files (since the <code>NETCDF3</code>
	3294	data model does not have unsigned integer data types).</p>
	3295
	3296	<p>The default value of <code><a href="#Variable.scale">scale</a></code> is <code>True</code>
	3297	(automatic conversions are performed).</p>
	3298	</div>
	3299
	3300
	3301	</div>
	3302	<div id="Variable.set_auto_mask" class="classattr">
	3303	<div class="attr function"><a class="headerlink" href="#Variable.set_auto_mask"># &nbsp</a>
	3304
	3305
	3306	<span class="def">def</span>
	3307	<span class="name">set_auto_mask</span><span class="signature">(unknown)</span>:
	3308	</div>
	3309
	3310
	3311	<div class="docstring"><p><strong><code>set_auto_mask(self,mask)</code></strong></p>
	3312
	3313	<p>turn on or off automatic conversion of variable data to and
	3314	from masked arrays .</p>
	3315
	3316	<p>If <code><a href="#Variable.mask">mask</a></code> is set to <code>True</code>, when data is read from a variable
	3317	it is converted to a masked array if any of the values are exactly
	3318	equal to the either the netCDF _FillValue or the value specified by the
	3319	missing_value variable attribute. The fill_value of the masked array
	3320	is set to the missing_value attribute (if it exists), otherwise
	3321	the netCDF _FillValue attribute (which has a default value
	3322	for each data type). If the variable has no missing_value attribute, the
	3323	_FillValue is used instead. If the variable has valid_min/valid_max and
	3324	missing_value attributes, data outside the specified range will be masked.
	3325	When data is written to a variable, the masked
	3326	array is converted back to a regular numpy array by replacing all the
	3327	masked values by the missing_value attribute of the variable (if it
	3328	exists). If the variable has no missing_value attribute, the _FillValue
	3329	is used instead. </p>
	3330
	3331	<p>The default value of <code><a href="#Variable.mask">mask</a></code> is <code>True</code>
	3332	(automatic conversions are performed).</p>
	3333	</div>
	3334
	3335
	3336	</div>
	3337	<div id="Variable.set_always_mask" class="classattr">
	3338	<div class="attr function"><a class="headerlink" href="#Variable.set_always_mask"># &nbsp</a>
	3339
	3340
	3341	<span class="def">def</span>
	3342	<span class="name">set_always_mask</span><span class="signature">(unknown)</span>:
	3343	</div>
	3344
	3345
	3346	<div class="docstring"><p><strong><code>set_always_mask(self,always_mask)</code></strong></p>
	3347
	3348	<p>turn on or off conversion of data without missing values to regular
	3349	numpy arrays.</p>
	3350
	3351	<p><code><a href="#Variable.always_mask">always_mask</a></code> is a Boolean determining if automatic conversion of
	3352	masked arrays with no missing values to regular numpy arrays shall be
	3353	applied. Default is True. Set to False to restore the default behaviour
	3354	in versions prior to 1.4.1 (numpy array returned unless missing values are present,
	3355	otherwise masked array returned).</p>
	3356	</div>
	3357
	3358
	3359	</div>
	3360	<div id="Variable.set_ncstring_attrs" class="classattr">
	3361	<div class="attr function"><a class="headerlink" href="#Variable.set_ncstring_attrs"># &nbsp</a>
	3362
	3363
	3364	<span class="def">def</span>
	3365	<span class="name">set_ncstring_attrs</span><span class="signature">(unknown)</span>:
	3366	</div>
	3367
	3368
	3369	<div class="docstring"><p><strong><code>set_always_mask(self,ncstring_attrs)</code></strong></p>
	3370
	3371	<p>turn on or off creating NC_STRING string attributes.</p>
	3372
	3373	<p>If <code>ncstring_attrs</code> is set to <code>True</code> then text attributes will be variable-length
	3374	NC_STRINGs.</p>
	3375
	3376	<p>The default value of <code>ncstring_attrs</code> is <code>False</code> (writing ascii text attributes as
	3377	NC_CHAR).</p>
	3378	</div>
	3379
	3380
	3381	</div>
	3382	<div id="Variable.set_collective" class="classattr">
	3383	<div class="attr function"><a class="headerlink" href="#Variable.set_collective"># &nbsp</a>
	3384
	3385
	3386	<span class="def">def</span>
	3387	<span class="name">set_collective</span><span class="signature">(unknown)</span>:
	3388	</div>
	3389
	3390
	3391	<div class="docstring"><p><strong><code>set_collective(self,True_or_False)</code></strong></p>
	3392
	3393	<p>turn on or off collective parallel IO access. Ignored if file is not
	3394	open for parallel access.</p>
	3395	</div>
	3396
	3397
	3398	</div>
	3399	<div id="Variable.get_dims" class="classattr">
	3400	<div class="attr function"><a class="headerlink" href="#Variable.get_dims"># &nbsp</a>
	3401
	3402
	3403	<span class="def">def</span>
	3404	<span class="name">get_dims</span><span class="signature">(unknown)</span>:
	3405	</div>
	3406
	3407
	3408	<div class="docstring"><p><strong><code>get_dims(self)</code></strong></p>
	3409
	3410	<p>return a tuple of <code><a href="#Dimension">Dimension</a></code> instances associated with this
	3411	<code><a href="#Variable">Variable</a></code>.</p>
	3412	</div>
	3413
	3414
	3415	</div>
	3416	<div id="Variable.name" class="classattr">
	3417	<div class="attr variable"><a class="headerlink" href="#Variable.name"># &nbsp</a>
	3418
	3419	<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3420	</div>
	3421
	3422	<div class="docstring"><p>string name of Variable instance</p>
	3423	</div>
	3424
	3425
	3426	</div>
	3427	<div id="Variable.datatype" class="classattr">
	3428	<div class="attr variable"><a class="headerlink" href="#Variable.datatype"># &nbsp</a>
	3429
	3430	<span class="name">datatype</span><span class="default_value"> = <attribute 'datatype' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3431	</div>
	3432
	3433	<div class="docstring"><p>numpy data type (for primitive data types) or
	3434	VLType/CompoundType/EnumType instance
	3435	(for compound, vlen or enum data types)</p>
	3436	</div>
	3437
	3438
	3439	</div>
	3440	<div id="Variable.shape" class="classattr">
	3441	<div class="attr variable"><a class="headerlink" href="#Variable.shape"># &nbsp</a>
	3442
	3443	<span class="name">shape</span><span class="default_value"> = <attribute 'shape' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3444	</div>
	3445
	3446	<div class="docstring"><p>find current sizes of all variable dimensions</p>
	3447	</div>
	3448
	3449
	3450	</div>
	3451	<div id="Variable.size" class="classattr">
	3452	<div class="attr variable"><a class="headerlink" href="#Variable.size"># &nbsp</a>
	3453
	3454	<span class="name">size</span><span class="default_value"> = <attribute 'size' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3455	</div>
	3456
	3457	<div class="docstring"><p>Return the number of stored elements.</p>
	3458	</div>
	3459
	3460
	3461	</div>
	3462	<div id="Variable.dimensions" class="classattr">
	3463	<div class="attr variable"><a class="headerlink" href="#Variable.dimensions"># &nbsp</a>
	3464
	3465	<span class="name">dimensions</span><span class="default_value"> = <attribute 'dimensions' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3466	</div>
	3467
	3468	<div class="docstring"><p>get variables's dimension names</p>
	3469	</div>
	3470
	3471
	3472	</div>
	3473	<div id="Variable.ndim" class="classattr">
	3474	<div class="attr variable"><a class="headerlink" href="#Variable.ndim"># &nbsp</a>
	3475
	3476	<span class="name">ndim</span><span class="default_value"> = <attribute 'ndim' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3477	</div>
	3478
	3479
	3480
	3481	</div>
	3482	<div id="Variable.dtype" class="classattr">
	3483	<div class="attr variable"><a class="headerlink" href="#Variable.dtype"># &nbsp</a>
	3484
	3485	<span class="name">dtype</span><span class="default_value"> = <attribute 'dtype' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3486	</div>
	3487
	3488
	3489
	3490	</div>
	3491	<div id="Variable.mask" class="classattr">
	3492	<div class="attr variable"><a class="headerlink" href="#Variable.mask"># &nbsp</a>
	3493
	3494	<span class="name">mask</span><span class="default_value"> = <attribute 'mask' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3495	</div>
	3496
	3497
	3498
	3499	</div>
	3500	<div id="Variable.scale" class="classattr">
	3501	<div class="attr variable"><a class="headerlink" href="#Variable.scale"># &nbsp</a>
	3502
	3503	<span class="name">scale</span><span class="default_value"> = <attribute 'scale' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3504	</div>
	3505
	3506
	3507
	3508	</div>
	3509	<div id="Variable.always_mask" class="classattr">
	3510	<div class="attr variable"><a class="headerlink" href="#Variable.always_mask"># &nbsp</a>
	3511
	3512	<span class="name">always_mask</span><span class="default_value"> = <attribute 'always_mask' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3513	</div>
	3514
	3515
	3516
	3517	</div>
	3518	<div id="Variable.chartostring" class="classattr">
	3519	<div class="attr variable"><a class="headerlink" href="#Variable.chartostring"># &nbsp</a>
	3520
	3521	<span class="name">chartostring</span><span class="default_value"> = <attribute 'chartostring' of '<a href="#_netCDF4.Variable">netCDF4._netCDF4.Variable</a>' objects></span>
	3522	</div>
	3523
	3524
	3525
	3526	</div>
	3527	</section>
	3528	<section id="Dimension">
	3529	<div class="attr class">
	3530	<a class="headerlink" href="#Dimension"># &nbsp</a>
	3531
	3532
	3533	<span class="def">class</span>
	3534	<span class="name">Dimension</span>:
	3535	</div>
	3536
	3537
	3538	<div class="docstring"><p>A netCDF <code><a href="#Dimension">Dimension</a></code> is used to describe the coordinates of a <code><a href="#Variable">Variable</a></code>.
	3539	See <code><a href="#Dimension.__init__">Dimension.__init__</a></code> for more details.</p>
	3540
	3541	<p>The current maximum size of a <code><a href="#Dimension">Dimension</a></code> instance can be obtained by
	3542	calling the python <code>len</code> function on the <code><a href="#Dimension">Dimension</a></code> instance. The
	3543	<code><a href="#Dimension.isunlimited">Dimension.isunlimited</a></code> method of a <code><a href="#Dimension">Dimension</a></code> instance can be used to
	3544	determine if the dimension is unlimited.</p>
	3545
	3546	<p>Read-only class variables:</p>
	3547
	3548	<p><strong><code><a href="#Dimension.name">name</a></code></strong>: String name, used when creating a <code><a href="#Variable">Variable</a></code> with
	3549	<code><a href="#Dataset.createVariable">Dataset.createVariable</a></code>.</p>
	3550
	3551	<p><strong><code><a href="#Dimension.size">size</a></code></strong>: Current <code><a href="#Dimension">Dimension</a></code> size (same as <code>len(d)</code>, where <code>d</code> is a
	3552	<code><a href="#Dimension">Dimension</a></code> instance).</p>
	3553	</div>
	3554
	3555
	3556	<div id="Dimension.__init__" class="classattr">
	3557	<div class="attr function"><a class="headerlink" href="#Dimension.__init__"># &nbsp</a>
	3558
	3559
	3560	<span class="name">Dimension</span><span class="signature">()</span>
	3561	</div>
	3562
	3563
	3564	<div class="docstring"><p><strong><code>__init__(self, group, name, size=None)</code></strong></p>
	3565
	3566	<p><code><a href="#Dimension">Dimension</a></code> constructor.</p>
	3567
	3568	<p><strong><code><a href="#Dimension.group">group</a></code></strong>: <code><a href="#Group">Group</a></code> instance to associate with dimension.</p>
	3569
	3570	<p><strong><code><a href="#Dimension.name">name</a></code></strong>: Name of the dimension.</p>
	3571
	3572	<p><strong><code><a href="#Dimension.size">size</a></code></strong>: Size of the dimension. <code>None</code> or 0 means unlimited. (Default <code>None</code>).</p>
	3573
	3574	<p><strong><em>Note</em></strong>: <code><a href="#Dimension">Dimension</a></code> instances should be created using the
	3575	<code><a href="#Dataset.createDimension">Dataset.createDimension</a></code> method of a <code><a href="#Group">Group</a></code> or
	3576	<code><a href="#Dataset">Dataset</a></code> instance, not using <code><a href="#Dimension.__init__">Dimension.__init__</a></code> directly.</p>
	3577	</div>
	3578
	3579
	3580	</div>
	3581	<div id="Dimension.group" class="classattr">
	3582	<div class="attr function"><a class="headerlink" href="#Dimension.group"># &nbsp</a>
	3583
	3584
	3585	<span class="def">def</span>
	3586	<span class="name">group</span><span class="signature">(unknown)</span>:
	3587	</div>
	3588
	3589
	3590	<div class="docstring"><p><strong><code>group(self)</code></strong></p>
	3591
	3592	<p>return the group that this <code><a href="#Dimension">Dimension</a></code> is a member of.</p>
	3593	</div>
	3594
	3595
	3596	</div>
	3597	<div id="Dimension.isunlimited" class="classattr">
	3598	<div class="attr function"><a class="headerlink" href="#Dimension.isunlimited"># &nbsp</a>
	3599
	3600
	3601	<span class="def">def</span>
	3602	<span class="name">isunlimited</span><span class="signature">(unknown)</span>:
	3603	</div>
	3604
	3605
	3606	<div class="docstring"><p><strong><code>isunlimited(self)</code></strong></p>
	3607
	3608	<p>returns <code>True</code> if the <code><a href="#Dimension">Dimension</a></code> instance is unlimited, <code>False</code> otherwise.</p>
	3609	</div>
	3610
	3611
	3612	</div>
	3613	<div id="Dimension.name" class="classattr">
	3614	<div class="attr variable"><a class="headerlink" href="#Dimension.name"># &nbsp</a>
	3615
	3616	<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#_netCDF4.Dimension">netCDF4._netCDF4.Dimension</a>' objects></span>
	3617	</div>
	3618
	3619	<div class="docstring"><p>string name of Dimension instance</p>
	3620	</div>
	3621
	3622
	3623	</div>
	3624	<div id="Dimension.size" class="classattr">
	3625	<div class="attr variable"><a class="headerlink" href="#Dimension.size"># &nbsp</a>
	3626
	3627	<span class="name">size</span><span class="default_value"> = <attribute 'size' of '<a href="#_netCDF4.Dimension">netCDF4._netCDF4.Dimension</a>' objects></span>
	3628	</div>
	3629
	3630	<div class="docstring"><p>current size of Dimension (calls <code>len</code> on Dimension instance)</p>
	3631	</div>
	3632
2701	3633
2702	3634	</div>
2703	3635	</section>

2707	3639
2708	3640
2709	3641	<span class="def">class</span>
2710		<span class="name">Group</span>(<span class="base"><a href="#Dataset">Dataset</a></span>):
	3642	<span class="name">Group</span>(<span class="base"><a href="#Dataset">netCDF4.Dataset</a></span>):
2711	3643	</div>
2712	3644
2713	3645
2714	3646	<div class="docstring"><p>Groups define a hierarchical namespace within a netCDF file. They are
2715		analogous to directories in a unix filesystem. Each <a href="#Group">Group</a> behaves like
2716		a <a href="#Dataset">Dataset</a> within a Dataset, and can contain it's own variables,
2717		dimensions and attributes (and other Groups). See <a href="#<a href="#Group.__init__">Group.__init__</a>"><a href="#Group.__init__">Group.__init__</a></a>
	3647	analogous to directories in a unix filesystem. Each <code><a href="#Group">Group</a></code> behaves like
	3648	a <code><a href="#Dataset">Dataset</a></code> within a Dataset, and can contain it's own variables,
	3649	dimensions and attributes (and other Groups). See <code><a href="#Group.__init__">Group.__init__</a></code>
2718	3650	for more details.</p>
2719	3651
2720		<p><a href="#Group">Group</a> inherits from <a href="#Dataset">Dataset</a>, so all the
2721		<a href="#Dataset">Dataset</a> class methods and variables are available
2722		to a <a href="#Group">Group</a> instance (except the <code>close</code> method).</p>
	3652	<p><code><a href="#Group">Group</a></code> inherits from <code><a href="#Dataset">Dataset</a></code>, so all the
	3653	<code><a href="#Dataset">Dataset</a></code> class methods and variables are available
	3654	to a <code><a href="#Group">Group</a></code> instance (except the <code><a href="#Group.close">close</a></code> method).</p>
2723	3655
2724	3656	<p>Additional read-only class variables:</p>
2725	3657
2726		<p><strong><code>name</code></strong>: String describing the group name.</p>
	3658	<p><strong><code><a href="#Group.name">name</a></code></strong>: String describing the group name.</p>
2727	3659	</div>
2728	3660
2729	3661

2731	3663	<div class="attr function"><a class="headerlink" href="#Group.__init__"># &nbsp</a>
2732	3664
2733	3665
2734		<span class="name">Group</span><span class="signature">(unknown)</span> </div>
	3666	<span class="name">Group</span><span class="signature">()</span>
	3667	</div>
2735	3668
2736	3669
2737	3670	<div class="docstring"><p><strong><code>__init__(self, parent, name)</code></strong>
2738		<a href="#Group">Group</a> constructor.</p>
2739
2740		<p><strong><code>parent</code></strong>: <a href="#Group">Group</a> instance for the parent group. If being created
2741		in the root group, use a <a href="#Dataset">Dataset</a> instance.</p>
2742
2743		<p><strong><code>name</code></strong>: - Name of the group.</p>
2744
2745		<p><strong><em>Note</em></strong>: <a href="#Group">Group</a> instances should be created using the
2746		<a href="#<a href="#Dataset.createGroup">Dataset.createGroup</a>"><a href="#Dataset.createGroup">Dataset.createGroup</a></a> method of a <a href="#Dataset">Dataset</a> instance, or
2747		another <a href="#Group">Group</a> instance, not using this class directly.</p>
	3671	<code><a href="#Group">Group</a></code> constructor.</p>
	3672
	3673	<p><strong><code><a href="#Group.parent">parent</a></code></strong>: <code><a href="#Group">Group</a></code> instance for the parent group. If being created
	3674	in the root group, use a <code><a href="#Dataset">Dataset</a></code> instance.</p>
	3675
	3676	<p><strong><code><a href="#Group.name">name</a></code></strong>: - Name of the group.</p>
	3677
	3678	<p><strong><em>Note</em></strong>: <code><a href="#Group">Group</a></code> instances should be created using the
	3679	<code><a href="#Dataset.createGroup">Dataset.createGroup</a></code> method of a <code><a href="#Dataset">Dataset</a></code> instance, or
	3680	another <code><a href="#Group">Group</a></code> instance, not using this class directly.</p>
2748	3681	</div>
2749	3682
2750	3683

2760	3693
2761	3694	<div class="docstring"><p><strong><code>close(self)</code></strong></p>
2762	3695
2763		<p>overrides <a href="#Dataset">Dataset</a> close method which does not apply to <a href="#Group">Group</a>
	3696	<p>overrides <code><a href="#Dataset">Dataset</a></code> close method which does not apply to <code><a href="#Group">Group</a></code>
2764	3697	instances, raises IOError.</p>
2765	3698	</div>
2766	3699

2769	3702	<div class="inherited">
2770	3703	<h5>Inherited Members</h5>
2771	3704	<dl>
2772		<div><dt><a href="#Dataset">netCDF4._netCDF4.Dataset</a></dt>
	3705	<div><dt><a href="#Dataset">Dataset</a></dt>
2773	3706	<dd id="Group.filepath" class="function"><a href="#Dataset.filepath">filepath</a></dd>
2774	3707	<dd id="Group.isopen" class="function"><a href="#Dataset.isopen">isopen</a></dd>
2775	3708	<dd id="Group.sync" class="function"><a href="#Dataset.sync">sync</a></dd>

2818	3751	</dl>
2819	3752	</div>
2820	3753	</section>
2821		<section id="Dimension">
2822		<div class="attr class">
2823		<a class="headerlink" href="#Dimension"># &nbsp</a>
2824
2825
2826		<span class="def">class</span>
2827		<span class="name">Dimension</span>:
2828		</div>
2829
2830
2831		<div class="docstring"><p>A netCDF <a href="#Dimension">Dimension</a> is used to describe the coordinates of a <a href="#Variable">Variable</a>.
2832		See <a href="#<a href="#Dimension.__init__">Dimension.__init__</a>"><a href="#Dimension.__init__">Dimension.__init__</a></a> for more details.</p>
2833
2834		<p>The current maximum size of a <a href="#Dimension">Dimension</a> instance can be obtained by
2835		calling the python <code>len</code> function on the <a href="#Dimension">Dimension</a> instance. The
2836		<a href="#<a href="#Dimension.isunlimited">Dimension.isunlimited</a>"><a href="#Dimension.isunlimited">Dimension.isunlimited</a></a> method of a <a href="#Dimension">Dimension</a> instance can be used to
2837		determine if the dimension is unlimited.</p>
2838
2839		<p>Read-only class variables:</p>
2840
2841		<p><strong><code>name</code></strong>: String name, used when creating a <a href="#Variable">Variable</a> with
2842		<a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a>.</p>
2843
2844		<p><strong><code>size</code></strong>: Current <a href="#Dimension">Dimension</a> size (same as <code>len(d)</code>, where <code>d</code> is a
2845		<a href="#Dimension">Dimension</a> instance).</p>
2846		</div>
2847
2848
2849		<div id="Dimension.__init__" class="classattr">
2850		<div class="attr function"><a class="headerlink" href="#Dimension.__init__"># &nbsp</a>
2851
2852
2853		<span class="name">Dimension</span><span class="signature">(unknown)</span> </div>
2854
2855
2856		<div class="docstring"><p><strong><code>__init__(self, group, name, size=None)</code></strong></p>
2857
2858		<p><a href="#Dimension">Dimension</a> constructor.</p>
2859
2860		<p><strong><code>group</code></strong>: <a href="#Group">Group</a> instance to associate with dimension.</p>
2861
2862		<p><strong><code>name</code></strong>: Name of the dimension.</p>
2863
2864		<p><strong><code>size</code></strong>: Size of the dimension. <code>None</code> or 0 means unlimited. (Default <code>None</code>).</p>
2865
2866		<p><strong><em>Note</em></strong>: <a href="#Dimension">Dimension</a> instances should be created using the
2867		<a href="#<a href="#Dataset.createDimension">Dataset.createDimension</a>"><a href="#Dataset.createDimension">Dataset.createDimension</a></a> method of a <a href="#Group">Group</a> or
2868		<a href="#Dataset">Dataset</a> instance, not using <a href="#<a href="#Dimension.__init__">Dimension.__init__</a>"><a href="#Dimension.__init__">Dimension.__init__</a></a> directly.</p>
2869		</div>
2870
2871
2872		</div>
2873		<div id="Dimension.group" class="classattr">
2874		<div class="attr function"><a class="headerlink" href="#Dimension.group"># &nbsp</a>
2875
2876
2877		<span class="def">def</span>
2878		<span class="name">group</span><span class="signature">(unknown)</span>:
2879		</div>
2880
2881
2882		<div class="docstring"><p><strong><code>group(self)</code></strong></p>
2883
2884		<p>return the group that this <a href="#Dimension">Dimension</a> is a member of.</p>
2885		</div>
2886
2887
2888		</div>
2889		<div id="Dimension.isunlimited" class="classattr">
2890		<div class="attr function"><a class="headerlink" href="#Dimension.isunlimited"># &nbsp</a>
2891
2892
2893		<span class="def">def</span>
2894		<span class="name">isunlimited</span><span class="signature">(unknown)</span>:
2895		</div>
2896
2897
2898		<div class="docstring"><p><strong><code>isunlimited(self)</code></strong></p>
2899
2900		<p>returns <code>True</code> if the <a href="#Dimension">Dimension</a> instance is unlimited, <code>False</code> otherwise.</p>
2901		</div>
2902
2903
2904		</div>
2905		<div id="Dimension.name" class="classattr">
2906		<div class="attr variable"><a class="headerlink" href="#Dimension.name"># &nbsp</a>
2907
2908		<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#Dimension">netCDF4._netCDF4.Dimension</a>' objects></span>
2909		</div>
2910
2911
2912
2913		</div>
2914		<div id="Dimension.size" class="classattr">
2915		<div class="attr variable"><a class="headerlink" href="#Dimension.size"># &nbsp</a>
2916
2917		<span class="name">size</span><span class="default_value"> = <attribute 'size' of '<a href="#Dimension">netCDF4._netCDF4.Dimension</a>' objects></span>
2918		</div>
2919
2920
2921
2922		</div>
2923		</section>
2924		<section id="Variable">
2925		<div class="attr class">
2926		<a class="headerlink" href="#Variable"># &nbsp</a>
2927
2928
2929		<span class="def">class</span>
2930		<span class="name">Variable</span>:
2931		</div>
2932
2933
2934		<div class="docstring"><p>A netCDF <a href="#Variable">Variable</a> is used to read and write netCDF data. They are
2935		analogous to numpy array objects. See <a href="#<a href="#Variable.__init__">Variable.__init__</a>"><a href="#Variable.__init__">Variable.__init__</a></a> for more
2936		details.</p>
2937
2938		<p>A list of attribute names corresponding to netCDF attributes defined for
2939		the variable can be obtained with the <a href="#<a href="#Variable.ncattrs">Variable.ncattrs</a>"><a href="#Variable.ncattrs">Variable.ncattrs</a></a> method. These
2940		attributes can be created by assigning to an attribute of the
2941		<a href="#Variable">Variable</a> instance. A dictionary containing all the netCDF attribute
2942		name/value pairs is provided by the <code>__dict__</code> attribute of a
2943		<a href="#Variable">Variable</a> instance.</p>
2944
2945		<p>The following class variables are read-only:</p>
2946
2947		<p><strong><code>dimensions</code></strong>: A tuple containing the names of the
2948		dimensions associated with this variable.</p>
2949
2950		<p><strong><code>dtype</code></strong>: A numpy dtype object describing the
2951		variable's data type.</p>
2952
2953		<p><strong><code>ndim</code></strong>: The number of variable dimensions.</p>
2954
2955		<p><strong><code>shape</code></strong>: A tuple with the current shape (length of all dimensions).</p>
2956
2957		<p><strong><code>scale</code></strong>: If True, <code>scale_factor</code> and <code>add_offset</code> are
2958		applied, and signed integer data is automatically converted to
2959		unsigned integer data if the <code>_Unsigned</code> attribute is set.
2960		Default is <code>True</code>, can be reset using <a href="#<a href="#Variable.set_auto_scale">Variable.set_auto_scale</a>"><a href="#Variable.set_auto_scale">Variable.set_auto_scale</a></a> and
2961		<a href="#<a href="#Variable.set_auto_maskandscale">Variable.set_auto_maskandscale</a>"><a href="#Variable.set_auto_maskandscale">Variable.set_auto_maskandscale</a></a> methods.</p>
2962
2963		<p><strong><code>mask</code></strong>: If True, data is automatically converted to/from masked
2964		arrays when missing values or fill values are present. Default is <code>True</code>, can be
2965		reset using <a href="#<a href="#Variable.set_auto_mask">Variable.set_auto_mask</a>"><a href="#Variable.set_auto_mask">Variable.set_auto_mask</a></a> and <a href="#<a href="#Variable.set_auto_maskandscale">Variable.set_auto_maskandscale</a>"><a href="#Variable.set_auto_maskandscale">Variable.set_auto_maskandscale</a></a>
2966		methods.</p>
2967
2968		<p><strong><code><a href="#chartostring">chartostring</a></code></strong>: If True, data is automatically converted to/from character
2969		arrays to string arrays when the <code>_Encoding</code> variable attribute is set.
2970		Default is <code>True</code>, can be reset using
2971		<a href="#<a href="#Variable.set_auto_chartostring">Variable.set_auto_chartostring</a>"><a href="#Variable.set_auto_chartostring">Variable.set_auto_chartostring</a></a> method.</p>
2972
2973		<p><strong><code>least_significant_digit</code></strong>: Describes the power of ten of the
2974		smallest decimal place in the data the contains a reliable value. Data is
2975		truncated to this decimal place when it is assigned to the <a href="#Variable">Variable</a>
2976		instance. If <code>None</code>, the data is not truncated.</p>
2977
2978		<p><strong><code>__orthogonal_indexing__</code></strong>: Always <code>True</code>. Indicates to client code
2979		that the object supports 'orthogonal indexing', which means that slices
2980		that are 1d arrays or lists slice along each dimension independently. This
2981		behavior is similar to Fortran or Matlab, but different than numpy.</p>
2982
2983		<p><strong><code>datatype</code></strong>: numpy data type (for primitive data types) or VLType/CompoundType
2984		instance (for compound or vlen data types).</p>
2985
2986		<p><strong><code>name</code></strong>: String name.</p>
2987
2988		<p><strong><code>size</code></strong>: The number of stored elements.</p>
2989		</div>
2990
2991
2992		<div id="Variable.__init__" class="classattr">
2993		<div class="attr function"><a class="headerlink" href="#Variable.__init__"># &nbsp</a>
2994
2995
2996		<span class="name">Variable</span><span class="signature">(unknown)</span> </div>
2997
2998
2999		<div class="docstring"><p><strong><code>__init__(self, group, name, datatype, dimensions=(), zlib=False,
3000		complevel=4, shuffle=True, fletcher32=False, contiguous=False,
3001		chunksizes=None, endian='native',
3002		least_significant_digit=None,fill_value=None,chunk_cache=None)</code></strong></p>
3003
3004		<p><a href="#Variable">Variable</a> constructor.</p>
3005
3006		<p><strong><code>group</code></strong>: <a href="#Group">Group</a> or <a href="#Dataset">Dataset</a> instance to associate with variable.</p>
3007
3008		<p><strong><code>name</code></strong>: Name of the variable.</p>
3009
3010		<p><strong><code>datatype</code></strong>: <a href="#Variable">Variable</a> data type. Can be specified by providing a
3011		numpy dtype object, or a string that describes a numpy dtype object.
3012		Supported values, corresponding to <code>str</code> attribute of numpy dtype
3013		objects, include <code>'f4'</code> (32-bit floating point), <code>'f8'</code> (64-bit floating
3014		point), <code>'i4'</code> (32-bit signed integer), <code>'i2'</code> (16-bit signed integer),
3015		<code>'i8'</code> (64-bit signed integer), <code>'i4'</code> (8-bit signed integer), <code>'i1'</code>
3016		(8-bit signed integer), <code>'u1'</code> (8-bit unsigned integer), <code>'u2'</code> (16-bit
3017		unsigned integer), <code>'u4'</code> (32-bit unsigned integer), <code>'u8'</code> (64-bit
3018		unsigned integer), or <code>'S1'</code> (single-character string). From
3019		compatibility with Scientific.IO.NetCDF, the old Numeric single character
3020		typecodes can also be used (<code>'f'</code> instead of <code>'f4'</code>, <code>'d'</code> instead of
3021		<code>'f8'</code>, <code>'h'</code> or <code>'s'</code> instead of <code>'i2'</code>, <code>'b'</code> or <code>'B'</code> instead of
3022		<code>'i1'</code>, <code>'c'</code> instead of <code>'S1'</code>, and <code>'i'</code> or <code>'l'</code> instead of
3023		<code>'i4'</code>). <code>datatype</code> can also be a <a href="#CompoundType">CompoundType</a> instance
3024		(for a structured, or compound array), a <a href="#VLType">VLType</a> instance
3025		(for a variable-length array), or the python <code>str</code> builtin
3026		(for a variable-length string array). Numpy string and unicode datatypes with
3027		length greater than one are aliases for <code>str</code>.</p>
3028
3029		<p><strong><code>dimensions</code></strong>: a tuple containing the variable's dimension names
3030		(defined previously with <code>createDimension</code>). Default is an empty tuple
3031		which means the variable is a scalar (and therefore has no dimensions).</p>
3032
3033		<p><strong><code>zlib</code></strong>: if <code>True</code>, data assigned to the <a href="#Variable">Variable</a>
3034		instance is compressed on disk. Default <code>False</code>.</p>
3035
3036		<p><strong><code>complevel</code></strong>: the level of zlib compression to use (1 is the fastest,
3037		but poorest compression, 9 is the slowest but best compression). Default 4.
3038		Ignored if <code>zlib=False</code>.</p>
3039
3040		<p><strong><code>shuffle</code></strong>: if <code>True</code>, the HDF5 shuffle filter is applied
3041		to improve compression. Default <code>True</code>. Ignored if <code>zlib=False</code>.</p>
3042
3043		<p><strong><code>fletcher32</code></strong>: if <code>True</code> (default <code>False</code>), the Fletcher32 checksum
3044		algorithm is used for error detection.</p>
3045
3046		<p><strong><code>contiguous</code></strong>: if <code>True</code> (default <code>False</code>), the variable data is
3047		stored contiguously on disk. Default <code>False</code>. Setting to <code>True</code> for
3048		a variable with an unlimited dimension will trigger an error.</p>
3049
3050		<p><strong><code>chunksizes</code></strong>: Can be used to specify the HDF5 chunksizes for each
3051		dimension of the variable. A detailed discussion of HDF chunking and I/O
3052		performance is available
3053		<a href="http://www.hdfgroup.org/HDF5/doc/H5.user/Chunking.html">here</a>.
3054		Basically, you want the chunk size for each dimension to match as
3055		closely as possible the size of the data block that users will read
3056		from the file. <code>chunksizes</code> cannot be set if <code>contiguous=True</code>.</p>
3057
3058		<p><strong><code>endian</code></strong>: Can be used to control whether the
3059		data is stored in little or big endian format on disk. Possible
3060		values are <code>little, big</code> or <code>native</code> (default). The library
3061		will automatically handle endian conversions when the data is read,
3062		but if the data is always going to be read on a computer with the
3063		opposite format as the one used to create the file, there may be
3064		some performance advantage to be gained by setting the endian-ness.
3065		For netCDF 3 files (that don't use HDF5), only <code>endian='native'</code> is allowed.</p>
3066
3067		<p>The <code>zlib, complevel, shuffle, fletcher32, contiguous</code> and <code>chunksizes</code>
3068		keywords are silently ignored for netCDF 3 files that do not use HDF5.</p>
3069
3070		<p><strong><code>least_significant_digit</code></strong>: If specified, variable data will be
3071		truncated (quantized). In conjunction with <code>zlib=True</code> this produces
3072		'lossy', but significantly more efficient compression. For example, if
3073		<code>least_significant_digit=1</code>, data will be quantized using
3074		around(scale<em>data)/scale, where scale = 2</em>*bits, and bits is determined
3075		so that a precision of 0.1 is retained (in this case bits=4). Default is
3076		<code>None</code>, or no quantization.</p>
3077
3078		<p><strong><code>fill_value</code></strong>: If specified, the default netCDF <code>_FillValue</code> (the
3079		value that the variable gets filled with before any data is written to it)
3080		is replaced with this value. If fill_value is set to <code>False</code>, then
3081		the variable is not pre-filled. The default netCDF fill values can be found
3082		in <a href="#default_fillvals">default_fillvals</a>.</p>
3083
3084		<p><strong><code>chunk_cache</code></strong>: If specified, sets the chunk cache size for this variable.
3085		Persists as long as Dataset is open. Use <a href="#set_var_chunk_cache">set_var_chunk_cache</a> to
3086		change it when Dataset is re-opened. </p>
3087
3088		<p><strong><em>Note</em></strong>: <a href="#Variable">Variable</a> instances should be created using the
3089		<a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a> method of a <a href="#Dataset">Dataset</a> or
3090		<a href="#Group">Group</a> instance, not using this class directly.</p>
3091		</div>
3092
3093
3094		</div>
3095		<div id="Variable.group" class="classattr">
3096		<div class="attr function"><a class="headerlink" href="#Variable.group"># &nbsp</a>
3097
3098
3099		<span class="def">def</span>
3100		<span class="name">group</span><span class="signature">(unknown)</span>:
3101		</div>
3102
3103
3104		<div class="docstring"><p><strong><code>group(self)</code></strong></p>
3105
3106		<p>return the group that this <a href="#Variable">Variable</a> is a member of.</p>
3107		</div>
3108
3109
3110		</div>
3111		<div id="Variable.ncattrs" class="classattr">
3112		<div class="attr function"><a class="headerlink" href="#Variable.ncattrs"># &nbsp</a>
3113
3114
3115		<span class="def">def</span>
3116		<span class="name">ncattrs</span><span class="signature">(unknown)</span>:
3117		</div>
3118
3119
3120		<div class="docstring"><p><strong><code>ncattrs(self)</code></strong></p>
3121
3122		<p>return netCDF attribute names for this <a href="#Variable">Variable</a> in a list.</p>
3123		</div>
3124
3125
3126		</div>
3127		<div id="Variable.setncattr" class="classattr">
3128		<div class="attr function"><a class="headerlink" href="#Variable.setncattr"># &nbsp</a>
3129
3130
3131		<span class="def">def</span>
3132		<span class="name">setncattr</span><span class="signature">(unknown)</span>:
3133		</div>
3134
3135
3136		<div class="docstring"><p><strong><code>setncattr(self,name,value)</code></strong></p>
3137
3138		<p>set a netCDF variable attribute using name,value pair. Use if you need to set a
3139		netCDF attribute with the same name as one of the reserved python
3140		attributes.</p>
3141		</div>
3142
3143
3144		</div>
3145		<div id="Variable.setncattr_string" class="classattr">
3146		<div class="attr function"><a class="headerlink" href="#Variable.setncattr_string"># &nbsp</a>
3147
3148
3149		<span class="def">def</span>
3150		<span class="name">setncattr_string</span><span class="signature">(unknown)</span>:
3151		</div>
3152
3153
3154		<div class="docstring"><p><strong><code>setncattr_string(self,name,value)</code></strong></p>
3155
3156		<p>set a netCDF variable string attribute using name,value pair.
3157		Use if you need to ensure that a netCDF attribute is created with type
3158		<code>NC_STRING</code> if the file format is <code>NETCDF4</code>.
3159		Use if you need to set an attribute to an array of variable-length strings.</p>
3160		</div>
3161
3162
3163		</div>
3164		<div id="Variable.setncatts" class="classattr">
3165		<div class="attr function"><a class="headerlink" href="#Variable.setncatts"># &nbsp</a>
3166
3167
3168		<span class="def">def</span>
3169		<span class="name">setncatts</span><span class="signature">(unknown)</span>:
3170		</div>
3171
3172
3173		<div class="docstring"><p><strong><code>setncatts(self,attdict)</code></strong></p>
3174
3175		<p>set a bunch of netCDF variable attributes at once using a python dictionary.
3176		This may be faster when setting a lot of attributes for a <code>NETCDF3</code>
3177		formatted file, since nc_redef/nc_enddef is not called in between setting
3178		each attribute</p>
3179		</div>
3180
3181
3182		</div>
3183		<div id="Variable.getncattr" class="classattr">
3184		<div class="attr function"><a class="headerlink" href="#Variable.getncattr"># &nbsp</a>
3185
3186
3187		<span class="def">def</span>
3188		<span class="name">getncattr</span><span class="signature">(unknown)</span>:
3189		</div>
3190
3191
3192		<div class="docstring"><p><strong><code>getncattr(self,name)</code></strong></p>
3193
3194		<p>retrieve a netCDF variable attribute. Use if you need to set a
3195		netCDF attribute with the same name as one of the reserved python
3196		attributes.</p>
3197
3198		<p>option kwarg <code>encoding</code> can be used to specify the
3199		character encoding of a string attribute (default is <code>utf-8</code>).</p>
3200		</div>
3201
3202
3203		</div>
3204		<div id="Variable.delncattr" class="classattr">
3205		<div class="attr function"><a class="headerlink" href="#Variable.delncattr"># &nbsp</a>
3206
3207
3208		<span class="def">def</span>
3209		<span class="name">delncattr</span><span class="signature">(unknown)</span>:
3210		</div>
3211
3212
3213		<div class="docstring"><p><strong><code>delncattr(self,name,value)</code></strong></p>
3214
3215		<p>delete a netCDF variable attribute. Use if you need to delete a
3216		netCDF attribute with the same name as one of the reserved python
3217		attributes.</p>
3218		</div>
3219
3220
3221		</div>
3222		<div id="Variable.filters" class="classattr">
3223		<div class="attr function"><a class="headerlink" href="#Variable.filters"># &nbsp</a>
3224
3225
3226		<span class="def">def</span>
3227		<span class="name">filters</span><span class="signature">(unknown)</span>:
3228		</div>
3229
3230
3231		<div class="docstring"><p><strong><code>filters(self)</code></strong></p>
3232
3233		<p>return dictionary containing HDF5 filter parameters.</p>
3234		</div>
3235
3236
3237		</div>
3238		<div id="Variable.endian" class="classattr">
3239		<div class="attr function"><a class="headerlink" href="#Variable.endian"># &nbsp</a>
3240
3241
3242		<span class="def">def</span>
3243		<span class="name">endian</span><span class="signature">(unknown)</span>:
3244		</div>
3245
3246
3247		<div class="docstring"><p><strong><code>endian(self)</code></strong></p>
3248
3249		<p>return endian-ness (<code>little,big,native</code>) of variable (as stored in HDF5 file).</p>
3250		</div>
3251
3252
3253		</div>
3254		<div id="Variable.chunking" class="classattr">
3255		<div class="attr function"><a class="headerlink" href="#Variable.chunking"># &nbsp</a>
3256
3257
3258		<span class="def">def</span>
3259		<span class="name">chunking</span><span class="signature">(unknown)</span>:
3260		</div>
3261
3262
3263		<div class="docstring"><p><strong><code>chunking(self)</code></strong></p>
3264
3265		<p>return variable chunking information. If the dataset is
3266		defined to be contiguous (and hence there is no chunking) the word 'contiguous'
3267		is returned. Otherwise, a sequence with the chunksize for
3268		each dimension is returned.</p>
3269		</div>
3270
3271
3272		</div>
3273		<div id="Variable.get_var_chunk_cache" class="classattr">
3274		<div class="attr function"><a class="headerlink" href="#Variable.get_var_chunk_cache"># &nbsp</a>
3275
3276
3277		<span class="def">def</span>
3278		<span class="name">get_var_chunk_cache</span><span class="signature">(unknown)</span>:
3279		</div>
3280
3281
3282		<div class="docstring"><p><strong><code>get_var_chunk_cache(self)</code></strong></p>
3283
3284		<p>return variable chunk cache information in a tuple (size,nelems,preemption).
3285		See netcdf C library documentation for <code>nc_get_var_chunk_cache</code> for
3286		details.</p>
3287		</div>
3288
3289
3290		</div>
3291		<div id="Variable.set_var_chunk_cache" class="classattr">
3292		<div class="attr function"><a class="headerlink" href="#Variable.set_var_chunk_cache"># &nbsp</a>
3293
3294
3295		<span class="def">def</span>
3296		<span class="name">set_var_chunk_cache</span><span class="signature">(unknown)</span>:
3297		</div>
3298
3299
3300		<div class="docstring"><p><strong><code>set_var_chunk_cache(self,size=None,nelems=None,preemption=None)</code></strong></p>
3301
3302		<p>change variable chunk cache settings.
3303		See netcdf C library documentation for <code>nc_set_var_chunk_cache</code> for
3304		details.</p>
3305		</div>
3306
3307
3308		</div>
3309		<div id="Variable.renameAttribute" class="classattr">
3310		<div class="attr function"><a class="headerlink" href="#Variable.renameAttribute"># &nbsp</a>
3311
3312
3313		<span class="def">def</span>
3314		<span class="name">renameAttribute</span><span class="signature">(unknown)</span>:
3315		</div>
3316
3317
3318		<div class="docstring"><p><strong><code>renameAttribute(self, oldname, newname)</code></strong></p>
3319
3320		<p>rename a <a href="#Variable">Variable</a> attribute named <code>oldname</code> to <code>newname</code>.</p>
3321		</div>
3322
3323
3324		</div>
3325		<div id="Variable.assignValue" class="classattr">
3326		<div class="attr function"><a class="headerlink" href="#Variable.assignValue"># &nbsp</a>
3327
3328
3329		<span class="def">def</span>
3330		<span class="name">assignValue</span><span class="signature">(unknown)</span>:
3331		</div>
3332
3333
3334		<div class="docstring"><p><strong><code>assignValue(self, val)</code></strong></p>
3335
3336		<p>assign a value to a scalar variable. Provided for compatibility with
3337		Scientific.IO.NetCDF, can also be done by assigning to an Ellipsis slice ([...]).</p>
3338		</div>
3339
3340
3341		</div>
3342		<div id="Variable.getValue" class="classattr">
3343		<div class="attr function"><a class="headerlink" href="#Variable.getValue"># &nbsp</a>
3344
3345
3346		<span class="def">def</span>
3347		<span class="name">getValue</span><span class="signature">(unknown)</span>:
3348		</div>
3349
3350
3351		<div class="docstring"><p><strong><code>getValue(self)</code></strong></p>
3352
3353		<p>get the value of a scalar variable. Provided for compatibility with
3354		Scientific.IO.NetCDF, can also be done by slicing with an Ellipsis ([...]).</p>
3355		</div>
3356
3357
3358		</div>
3359		<div id="Variable.set_auto_chartostring" class="classattr">
3360		<div class="attr function"><a class="headerlink" href="#Variable.set_auto_chartostring"># &nbsp</a>
3361
3362
3363		<span class="def">def</span>
3364		<span class="name">set_auto_chartostring</span><span class="signature">(unknown)</span>:
3365		</div>
3366
3367
3368		<div class="docstring"><p><strong><code>set_auto_chartostring(self,chartostring)</code></strong></p>
3369
3370		<p>turn on or off automatic conversion of character variable data to and
3371		from numpy fixed length string arrays when the <code>_Encoding</code> variable attribute
3372		is set.</p>
3373
3374		<p>If <code><a href="#chartostring">chartostring</a></code> is set to <code>True</code>, when data is read from a character variable
3375		(dtype = <code>S1</code>) that has an <code>_Encoding</code> attribute, it is converted to a numpy
3376		fixed length unicode string array (dtype = <code>UN</code>, where <code>N</code> is the length
3377		of the the rightmost dimension of the variable). The value of <code>_Encoding</code>
3378		is the unicode encoding that is used to decode the bytes into strings.</p>
3379
3380		<p>When numpy string data is written to a variable it is converted back to
3381		indiviual bytes, with the number of bytes in each string equalling the
3382		rightmost dimension of the variable.</p>
3383
3384		<p>The default value of <code><a href="#chartostring">chartostring</a></code> is <code>True</code>
3385		(automatic conversions are performed).</p>
3386		</div>
3387
3388
3389		</div>
3390		<div id="Variable.use_nc_get_vars" class="classattr">
3391		<div class="attr function"><a class="headerlink" href="#Variable.use_nc_get_vars"># &nbsp</a>
3392
3393
3394		<span class="def">def</span>
3395		<span class="name">use_nc_get_vars</span><span class="signature">(unknown)</span>:
3396		</div>
3397
3398
3399		<div class="docstring"><p><strong><code>use_nc_get_vars(self,_use_get_vars)</code></strong></p>
3400
3401		<p>enable the use of netcdf library routine <code>nc_get_vars</code>
3402		to retrieve strided variable slices. By default,
3403		<code>nc_get_vars</code> may not used by default (depending on the
3404		version of the netcdf-c library being used) since it may be
3405		slower than multiple calls to the unstrided read routine <code>nc_get_vara</code>.</p>
3406		</div>
3407
3408
3409		</div>
3410		<div id="Variable.set_auto_maskandscale" class="classattr">
3411		<div class="attr function"><a class="headerlink" href="#Variable.set_auto_maskandscale"># &nbsp</a>
3412
3413
3414		<span class="def">def</span>
3415		<span class="name">set_auto_maskandscale</span><span class="signature">(unknown)</span>:
3416		</div>
3417
3418
3419		<div class="docstring"><p><strong><code>set_auto_maskandscale(self,maskandscale)</code></strong></p>
3420
3421		<p>turn on or off automatic conversion of variable data to and
3422		from masked arrays, automatic packing/unpacking of variable
3423		data using <code>scale_factor</code> and <code>add_offset</code> attributes and
3424		automatic conversion of signed integer data to unsigned integer
3425		data if the <code>_Unsigned</code> attribute exists.</p>
3426
3427		<p>If <code>maskandscale</code> is set to <code>True</code>, when data is read from a variable
3428		it is converted to a masked array if any of the values are exactly
3429		equal to the either the netCDF _FillValue or the value specified by the
3430		missing_value variable attribute. The fill_value of the masked array
3431		is set to the missing_value attribute (if it exists), otherwise
3432		the netCDF _FillValue attribute (which has a default value
3433		for each data type). When data is written to a variable, the masked
3434		array is converted back to a regular numpy array by replacing all the
3435		masked values by the missing_value attribute of the variable (if it
3436		exists). If the variable has no missing_value attribute, the _FillValue
3437		is used instead. If the variable has valid_min/valid_max and
3438		missing_value attributes, data outside the specified range will be
3439		set to missing_value.</p>
3440
3441		<p>If <code>maskandscale</code> is set to <code>True</code>, and the variable has a
3442		<code>scale_factor</code> or an <code>add_offset</code> attribute, then data read
3443		from that variable is unpacked using::</p>
3444
3445		<pre><code>data = self.scale_factor*data + self.add_offset
3446		</code></pre>
3447
3448		<p>When data is written to a variable it is packed using::</p>
3449
3450		<pre><code>data = (data - self.add_offset)/self.scale_factor
3451		</code></pre>
3452
3453		<p>If either scale_factor is present, but add_offset is missing, add_offset
3454		is assumed zero. If add_offset is present, but scale_factor is missing,
3455		scale_factor is assumed to be one.
3456		For more information on how <code>scale_factor</code> and <code>add_offset</code> can be
3457		used to provide simple compression, see the
3458		<a href="http://www.esrl.noaa.gov/psl/data/gridded/conventions/cdc_netcdf_standard.shtml">PSL metadata conventions</a>.</p>
3459
3460		<p>In addition, if <code>maskandscale</code> is set to <code>True</code>, and if the variable has an
3461		attribute <code>_Unsigned</code> set, and the variable has a signed integer data type,
3462		a view to the data is returned with the corresponding unsigned integer data type.
3463		This convention is used by the netcdf-java library to save unsigned integer
3464		data in <code>NETCDF3</code> or <code>NETCDF4_CLASSIC</code> files (since the <code>NETCDF3</code>
3465		data model does not have unsigned integer data types).</p>
3466
3467		<p>The default value of <code>maskandscale</code> is <code>True</code>
3468		(automatic conversions are performed).</p>
3469		</div>
3470
3471
3472		</div>
3473		<div id="Variable.set_auto_scale" class="classattr">
3474		<div class="attr function"><a class="headerlink" href="#Variable.set_auto_scale"># &nbsp</a>
3475
3476
3477		<span class="def">def</span>
3478		<span class="name">set_auto_scale</span><span class="signature">(unknown)</span>:
3479		</div>
3480
3481
3482		<div class="docstring"><p><strong><code>set_auto_scale(self,scale)</code></strong></p>
3483
3484		<p>turn on or off automatic packing/unpacking of variable
3485		data using <code>scale_factor</code> and <code>add_offset</code> attributes.
3486		Also turns on and off automatic conversion of signed integer data
3487		to unsigned integer data if the variable has an <code>_Unsigned</code>
3488		attribute.</p>
3489
3490		<p>If <code>scale</code> is set to <code>True</code>, and the variable has a
3491		<code>scale_factor</code> or an <code>add_offset</code> attribute, then data read
3492		from that variable is unpacked using::</p>
3493
3494		<pre><code>data = self.scale_factor*data + self.add_offset
3495		</code></pre>
3496
3497		<p>When data is written to a variable it is packed using::</p>
3498
3499		<pre><code>data = (data - self.add_offset)/self.scale_factor
3500		</code></pre>
3501
3502		<p>If either scale_factor is present, but add_offset is missing, add_offset
3503		is assumed zero. If add_offset is present, but scale_factor is missing,
3504		scale_factor is assumed to be one.
3505		For more information on how <code>scale_factor</code> and <code>add_offset</code> can be
3506		used to provide simple compression, see the
3507		<a href="http://www.esrl.noaa.gov/psl/data/gridded/conventions/cdc_netcdf_standard.shtml">PSL metadata conventions</a>.</p>
3508
3509		<p>In addition, if <code>scale</code> is set to <code>True</code>, and if the variable has an
3510		attribute <code>_Unsigned</code> set, and the variable has a signed integer data type,
3511		a view to the data is returned with the corresponding unsigned integer datatype.
3512		This convention is used by the netcdf-java library to save unsigned integer
3513		data in <code>NETCDF3</code> or <code>NETCDF4_CLASSIC</code> files (since the <code>NETCDF3</code>
3514		data model does not have unsigned integer data types).</p>
3515
3516		<p>The default value of <code>scale</code> is <code>True</code>
3517		(automatic conversions are performed).</p>
3518		</div>
3519
3520
3521		</div>
3522		<div id="Variable.set_auto_mask" class="classattr">
3523		<div class="attr function"><a class="headerlink" href="#Variable.set_auto_mask"># &nbsp</a>
3524
3525
3526		<span class="def">def</span>
3527		<span class="name">set_auto_mask</span><span class="signature">(unknown)</span>:
3528		</div>
3529
3530
3531		<div class="docstring"><p><strong><code>set_auto_mask(self,mask)</code></strong></p>
3532
3533		<p>turn on or off automatic conversion of variable data to and
3534		from masked arrays .</p>
3535
3536		<p>If <code>mask</code> is set to <code>True</code>, when data is read from a variable
3537		it is converted to a masked array if any of the values are exactly
3538		equal to the either the netCDF _FillValue or the value specified by the
3539		missing_value variable attribute. The fill_value of the masked array
3540		is set to the missing_value attribute (if it exists), otherwise
3541		the netCDF _FillValue attribute (which has a default value
3542		for each data type). When data is written to a variable, the masked
3543		array is converted back to a regular numpy array by replacing all the
3544		masked values by the missing_value attribute of the variable (if it
3545		exists). If the variable has no missing_value attribute, the _FillValue
3546		is used instead. If the variable has valid_min/valid_max and
3547		missing_value attributes, data outside the specified range will be
3548		set to missing_value.</p>
3549
3550		<p>The default value of <code>mask</code> is <code>True</code>
3551		(automatic conversions are performed).</p>
3552		</div>
3553
3554
3555		</div>
3556		<div id="Variable.set_always_mask" class="classattr">
3557		<div class="attr function"><a class="headerlink" href="#Variable.set_always_mask"># &nbsp</a>
3558
3559
3560		<span class="def">def</span>
3561		<span class="name">set_always_mask</span><span class="signature">(unknown)</span>:
3562		</div>
3563
3564
3565		<div class="docstring"><p><strong><code>set_always_mask(self,always_mask)</code></strong></p>
3566
3567		<p>turn on or off conversion of data without missing values to regular
3568		numpy arrays.</p>
3569
3570		<p><code>always_mask</code> is a Boolean determining if automatic conversion of
3571		masked arrays with no missing values to regular numpy arrays shall be
3572		applied. Default is True. Set to False to restore the default behaviour
3573		in versions prior to 1.4.1 (numpy array returned unless missing values are present,
3574		otherwise masked array returned).</p>
3575		</div>
3576
3577
3578		</div>
3579		<div id="Variable.set_ncstring_attrs" class="classattr">
3580		<div class="attr function"><a class="headerlink" href="#Variable.set_ncstring_attrs"># &nbsp</a>
3581
3582
3583		<span class="def">def</span>
3584		<span class="name">set_ncstring_attrs</span><span class="signature">(unknown)</span>:
3585		</div>
3586
3587
3588		<div class="docstring"><p><strong><code>set_always_mask(self,ncstring_attrs)</code></strong></p>
3589
3590		<p>turn on or off creating NC_STRING string attributes.</p>
3591
3592		<p>If <code>ncstring_attrs</code> is set to <code>True</code> then text attributes will be variable-length
3593		NC_STRINGs.</p>
3594
3595		<p>The default value of <code>ncstring_attrs</code> is <code>False</code> (writing ascii text attributes as
3596		NC_CHAR).</p>
3597		</div>
3598
3599
3600		</div>
3601		<div id="Variable.set_collective" class="classattr">
3602		<div class="attr function"><a class="headerlink" href="#Variable.set_collective"># &nbsp</a>
3603
3604
3605		<span class="def">def</span>
3606		<span class="name">set_collective</span><span class="signature">(unknown)</span>:
3607		</div>
3608
3609
3610		<div class="docstring"><p><strong><code>set_collective(self,True_or_False)</code></strong></p>
3611
3612		<p>turn on or off collective parallel IO access. Ignored if file is not
3613		open for parallel access.</p>
3614		</div>
3615
3616
3617		</div>
3618		<div id="Variable.get_dims" class="classattr">
3619		<div class="attr function"><a class="headerlink" href="#Variable.get_dims"># &nbsp</a>
3620
3621
3622		<span class="def">def</span>
3623		<span class="name">get_dims</span><span class="signature">(unknown)</span>:
3624		</div>
3625
3626
3627		<div class="docstring"><p><strong><code>get_dims(self)</code></strong></p>
3628
3629		<p>return a tuple of <a href="#Dimension">Dimension</a> instances associated with this
3630		<a href="#Variable">Variable</a>.</p>
3631		</div>
3632
3633
3634		</div>
3635		<div id="Variable.name" class="classattr">
3636		<div class="attr variable"><a class="headerlink" href="#Variable.name"># &nbsp</a>
3637
3638		<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3639		</div>
3640
3641
3642
3643		</div>
3644		<div id="Variable.datatype" class="classattr">
3645		<div class="attr variable"><a class="headerlink" href="#Variable.datatype"># &nbsp</a>
3646
3647		<span class="name">datatype</span><span class="default_value"> = <attribute 'datatype' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3648		</div>
3649
3650
3651
3652		</div>
3653		<div id="Variable.shape" class="classattr">
3654		<div class="attr variable"><a class="headerlink" href="#Variable.shape"># &nbsp</a>
3655
3656		<span class="name">shape</span><span class="default_value"> = <attribute 'shape' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3657		</div>
3658
3659
3660
3661		</div>
3662		<div id="Variable.size" class="classattr">
3663		<div class="attr variable"><a class="headerlink" href="#Variable.size"># &nbsp</a>
3664
3665		<span class="name">size</span><span class="default_value"> = <attribute 'size' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3666		</div>
3667
3668
3669
3670		</div>
3671		<div id="Variable.dimensions" class="classattr">
3672		<div class="attr variable"><a class="headerlink" href="#Variable.dimensions"># &nbsp</a>
3673
3674		<span class="name">dimensions</span><span class="default_value"> = <attribute 'dimensions' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3675		</div>
3676
3677
3678
3679		</div>
3680		<div id="Variable.ndim" class="classattr">
3681		<div class="attr variable"><a class="headerlink" href="#Variable.ndim"># &nbsp</a>
3682
3683		<span class="name">ndim</span><span class="default_value"> = <attribute 'ndim' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3684		</div>
3685
3686
3687
3688		</div>
3689		<div id="Variable.dtype" class="classattr">
3690		<div class="attr variable"><a class="headerlink" href="#Variable.dtype"># &nbsp</a>
3691
3692		<span class="name">dtype</span><span class="default_value"> = <attribute 'dtype' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3693		</div>
3694
3695
3696
3697		</div>
3698		<div id="Variable.mask" class="classattr">
3699		<div class="attr variable"><a class="headerlink" href="#Variable.mask"># &nbsp</a>
3700
3701		<span class="name">mask</span><span class="default_value"> = <attribute 'mask' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3702		</div>
3703
3704
3705
3706		</div>
3707		<div id="Variable.scale" class="classattr">
3708		<div class="attr variable"><a class="headerlink" href="#Variable.scale"># &nbsp</a>
3709
3710		<span class="name">scale</span><span class="default_value"> = <attribute 'scale' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3711		</div>
3712
3713
3714
3715		</div>
3716		<div id="Variable.always_mask" class="classattr">
3717		<div class="attr variable"><a class="headerlink" href="#Variable.always_mask"># &nbsp</a>
3718
3719		<span class="name">always_mask</span><span class="default_value"> = <attribute 'always_mask' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3720		</div>
3721
3722
3723
3724		</div>
3725		<div id="Variable.chartostring" class="classattr">
3726		<div class="attr variable"><a class="headerlink" href="#Variable.chartostring"># &nbsp</a>
3727
3728		<span class="name">chartostring</span><span class="default_value"> = <attribute 'chartostring' of '<a href="#Variable">netCDF4._netCDF4.Variable</a>' objects></span>
3729		</div>
3730
3731
3732
3733		</div>
3734		</section>
3735		<section id="CompoundType">
3736		<div class="attr class">
3737		<a class="headerlink" href="#CompoundType"># &nbsp</a>
3738
3739
3740		<span class="def">class</span>
3741		<span class="name">CompoundType</span>:
3742		</div>
3743
3744
3745		<div class="docstring"><p>A <a href="#CompoundType">CompoundType</a> instance is used to describe a compound data
3746		type, and can be passed to the the <a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a> method of
3747		a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> instance.
3748		Compound data types map to numpy structured arrays.
3749		See <a href="#<a href="#CompoundType.__init__">CompoundType.__init__</a>"><a href="#CompoundType.__init__">CompoundType.__init__</a></a> for more details.</p>
3750
3751		<p>The instance variables <code>dtype</code> and <code>name</code> should not be modified by
3752		the user.</p>
3753		</div>
3754
3755
3756		<div id="CompoundType.__init__" class="classattr">
3757		<div class="attr function"><a class="headerlink" href="#CompoundType.__init__"># &nbsp</a>
3758
3759
3760		<span class="name">CompoundType</span><span class="signature">(unknown)</span> </div>
3761
3762
3763		<div class="docstring"><p><strong><em><code>__init__(group, datatype, datatype_name)</code></em></strong></p>
3764
3765		<p>CompoundType constructor.</p>
3766
3767		<p><strong><code>group</code></strong>: <a href="#Group">Group</a> instance to associate with the compound datatype.</p>
3768
3769		<p><strong><code>datatype</code></strong>: A numpy dtype object describing a structured (a.k.a record)
3770		array. Can be composed of homogeneous numeric or character data types, or
3771		other structured array data types.</p>
3772
3773		<p><strong><code>datatype_name</code></strong>: a Python string containing a description of the
3774		compound data type.</p>
3775
3776		<p><strong><em>Note 1</em></strong>: When creating nested compound data types,
3777		the inner compound data types must already be associated with CompoundType
3778		instances (so create CompoundType instances for the innermost structures
3779		first).</p>
3780
3781		<p><strong><em>Note 2</em></strong>: <a href="#CompoundType">CompoundType</a> instances should be created using the
3782		<a href="#<a href="#Dataset.createCompoundType">Dataset.createCompoundType</a>"><a href="#Dataset.createCompoundType">Dataset.createCompoundType</a></a>
3783		method of a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> instance, not using this class directly.</p>
3784		</div>
3785
3786
3787		</div>
3788		<div id="CompoundType.dtype" class="classattr">
3789		<div class="attr variable"><a class="headerlink" href="#CompoundType.dtype"># &nbsp</a>
3790
3791		<span class="name">dtype</span><span class="default_value"> = <attribute 'dtype' of '<a href="#CompoundType">netCDF4._netCDF4.CompoundType</a>' objects></span>
3792		</div>
3793
3794
3795
3796		</div>
3797		<div id="CompoundType.dtype_view" class="classattr">
3798		<div class="attr variable"><a class="headerlink" href="#CompoundType.dtype_view"># &nbsp</a>
3799
3800		<span class="name">dtype_view</span><span class="default_value"> = <attribute 'dtype_view' of '<a href="#CompoundType">netCDF4._netCDF4.CompoundType</a>' objects></span>
3801		</div>
3802
3803
3804
3805		</div>
3806		<div id="CompoundType.name" class="classattr">
3807		<div class="attr variable"><a class="headerlink" href="#CompoundType.name"># &nbsp</a>
3808
3809		<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#CompoundType">netCDF4._netCDF4.CompoundType</a>' objects></span>
3810		</div>
3811
3812
3813
3814		</div>
3815		</section>
3816		<section id="VLType">
3817		<div class="attr class">
3818		<a class="headerlink" href="#VLType"># &nbsp</a>
3819
3820
3821		<span class="def">class</span>
3822		<span class="name">VLType</span>:
3823		</div>
3824
3825
3826		<div class="docstring"><p>A <a href="#VLType">VLType</a> instance is used to describe a variable length (VLEN) data
3827		type, and can be passed to the the <a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a> method of
3828		a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> instance. See
3829		<a href="#<a href="#VLType.__init__">VLType.__init__</a>"><a href="#VLType.__init__">VLType.__init__</a></a>for more details.</p>
3830
3831		<p>The instance variables <code>dtype</code> and <code>name</code> should not be modified by
3832		the user.</p>
3833		</div>
3834
3835
3836		<div id="VLType.__init__" class="classattr">
3837		<div class="attr function"><a class="headerlink" href="#VLType.__init__"># &nbsp</a>
3838
3839
3840		<span class="name">VLType</span><span class="signature">(unknown)</span> </div>
3841
3842
3843		<div class="docstring"><p><strong><code>__init__(group, datatype, datatype_name)</code></strong></p>
3844
3845		<p>VLType constructor.</p>
3846
3847		<p><strong><code>group</code></strong>: <a href="#Group">Group</a> instance to associate with the VLEN datatype.</p>
3848
3849		<p><strong><code>datatype</code></strong>: An numpy dtype object describing the component type for the
3850		variable length array.</p>
3851
3852		<p><strong><code>datatype_name</code></strong>: a Python string containing a description of the
3853		VLEN data type.</p>
3854
3855		<p><strong><em><code>Note</code></em></strong>: <a href="#VLType">VLType</a> instances should be created using the
3856		<a href="#<a href="#Dataset.createVLType">Dataset.createVLType</a>"><a href="#Dataset.createVLType">Dataset.createVLType</a></a>
3857		method of a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> instance, not using this class directly.</p>
3858		</div>
3859
3860
3861		</div>
3862		<div id="VLType.dtype" class="classattr">
3863		<div class="attr variable"><a class="headerlink" href="#VLType.dtype"># &nbsp</a>
3864
3865		<span class="name">dtype</span><span class="default_value"> = <attribute 'dtype' of '<a href="#VLType">netCDF4._netCDF4.VLType</a>' objects></span>
3866		</div>
3867
3868
3869
3870		</div>
3871		<div id="VLType.name" class="classattr">
3872		<div class="attr variable"><a class="headerlink" href="#VLType.name"># &nbsp</a>
3873
3874		<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#VLType">netCDF4._netCDF4.VLType</a>' objects></span>
3875		</div>
3876
3877
3878
3879		</div>
3880		</section>
3881		<section id="EnumType">
3882		<div class="attr class">
3883		<a class="headerlink" href="#EnumType"># &nbsp</a>
3884
3885
3886		<span class="def">class</span>
3887		<span class="name">EnumType</span>:
3888		</div>
3889
3890
3891		<div class="docstring"><p>A <a href="#EnumType">EnumType</a> instance is used to describe an Enum data
3892		type, and can be passed to the the <a href="#<a href="#Dataset.createVariable">Dataset.createVariable</a>"><a href="#Dataset.createVariable">Dataset.createVariable</a></a> method of
3893		a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> instance. See
3894		<a href="#<a href="#EnumType.__init__">EnumType.__init__</a>"><a href="#EnumType.__init__">EnumType.__init__</a></a> for more details.</p>
3895
3896		<p>The instance variables <code>dtype</code>, <code>name</code> and <code>enum_dict</code> should not be modified by
3897		the user.</p>
3898		</div>
3899
3900
3901		<div id="EnumType.__init__" class="classattr">
3902		<div class="attr function"><a class="headerlink" href="#EnumType.__init__"># &nbsp</a>
3903
3904
3905		<span class="name">EnumType</span><span class="signature">(unknown)</span> </div>
3906
3907
3908		<div class="docstring"><p><strong><code>__init__(group, datatype, datatype_name, enum_dict)</code></strong></p>
3909
3910		<p>EnumType constructor.</p>
3911
3912		<p><strong><code>group</code></strong>: <a href="#Group">Group</a> instance to associate with the VLEN datatype.</p>
3913
3914		<p><strong><code>datatype</code></strong>: An numpy integer dtype object describing the base type
3915		for the Enum.</p>
3916
3917		<p><strong><code>datatype_name</code></strong>: a Python string containing a description of the
3918		Enum data type.</p>
3919
3920		<p><strong><code>enum_dict</code></strong>: a Python dictionary containing the Enum field/value
3921		pairs.</p>
3922
3923		<p><strong><em><code>Note</code></em></strong>: <a href="#EnumType">EnumType</a> instances should be created using the
3924		<a href="#<a href="#Dataset.createEnumType">Dataset.createEnumType</a>"><a href="#Dataset.createEnumType">Dataset.createEnumType</a></a>
3925		method of a <a href="#Dataset">Dataset</a> or <a href="#Group">Group</a> instance, not using this class directly.</p>
3926		</div>
3927
3928
3929		</div>
3930		<div id="EnumType.dtype" class="classattr">
3931		<div class="attr variable"><a class="headerlink" href="#EnumType.dtype"># &nbsp</a>
3932
3933		<span class="name">dtype</span><span class="default_value"> = <attribute 'dtype' of '<a href="#EnumType">netCDF4._netCDF4.EnumType</a>' objects></span>
3934		</div>
3935
3936
3937
3938		</div>
3939		<div id="EnumType.name" class="classattr">
3940		<div class="attr variable"><a class="headerlink" href="#EnumType.name"># &nbsp</a>
3941
3942		<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#EnumType">netCDF4._netCDF4.EnumType</a>' objects></span>
3943		</div>
3944
3945
3946
3947		</div>
3948		<div id="EnumType.enum_dict" class="classattr">
3949		<div class="attr variable"><a class="headerlink" href="#EnumType.enum_dict"># &nbsp</a>
3950
3951		<span class="name">enum_dict</span><span class="default_value"> = <attribute 'enum_dict' of '<a href="#EnumType">netCDF4._netCDF4.EnumType</a>' objects></span>
3952		</div>
3953
3954
3955
3956		</div>
3957		</section>
3958		<section id="getlibversion">
3959		<div class="attr function"><a class="headerlink" href="#getlibversion"># &nbsp</a>
3960
3961
3962		<span class="def">def</span>
3963		<span class="name">getlibversion</span><span class="signature">(unknown)</span>:
3964		</div>
3965
3966
3967		<div class="docstring"><p><strong><code><a href="#getlibversion">getlibversion()</a></code></strong></p>
3968
3969		<p>returns a string describing the version of the netcdf library
3970		used to build the module, and when it was built.</p>
3971		</div>
3972
3973
3974		</section>
3975		<section id="get_chunk_cache">
3976		<div class="attr function"><a class="headerlink" href="#get_chunk_cache"># &nbsp</a>
3977
3978
3979		<span class="def">def</span>
3980		<span class="name">get_chunk_cache</span><span class="signature">(unknown)</span>:
3981		</div>
3982
3983
3984		<div class="docstring"><p><strong><code><a href="#get_chunk_cache">get_chunk_cache()</a></code></strong></p>
3985
3986		<p>return current netCDF chunk cache information in a tuple (size,nelems,preemption).
3987		See netcdf C library documentation for <code>nc_get_chunk_cache</code> for
3988		details. Values can be reset with <a href="#set_chunk_cache">set_chunk_cache</a>.</p>
3989		</div>
3990
3991
3992		</section>
3993		<section id="set_chunk_cache">
3994		<div class="attr function"><a class="headerlink" href="#set_chunk_cache"># &nbsp</a>
3995
3996
3997		<span class="def">def</span>
3998		<span class="name">set_chunk_cache</span><span class="signature">(unknown)</span>:
3999		</div>
4000
4001
4002		<div class="docstring"><p><strong><code>set_chunk_cache(self,size=None,nelems=None,preemption=None)</code></strong></p>
4003
4004		<p>change netCDF4 chunk cache settings.
4005		See netcdf C library documentation for <code>nc_set_chunk_cache</code> for
4006		details.</p>
4007		</div>
4008
4009
4010		</section>
4011		<section id="stringtoarr">
4012		<div class="attr function"><a class="headerlink" href="#stringtoarr"># &nbsp</a>
4013
4014
4015		<span class="def">def</span>
4016		<span class="name">stringtoarr</span><span class="signature">(unknown)</span>:
4017		</div>
4018
4019
4020		<div class="docstring"><p><strong><code>stringtoarr(a, NUMCHARS,dtype='S')</code></strong></p>
4021
4022		<p>convert a string to a character array of length <code>NUMCHARS</code></p>
4023
4024		<p><strong><code>a</code></strong>: Input python string.</p>
4025
4026		<p><strong><code>NUMCHARS</code></strong>: number of characters used to represent string
4027		(if len(a) < <code>NUMCHARS</code>, it will be padded on the right with blanks).</p>
4028
4029		<p><strong><code>dtype</code></strong>: type of numpy array to return. Default is <code>'S'</code>, which
4030		means an array of dtype <code>'S1'</code> will be returned. If dtype=<code>'U'</code>, a
4031		unicode array (dtype = <code>'U1'</code>) will be returned.</p>
4032
4033		<p>returns a rank 1 numpy character array of length NUMCHARS with datatype <code>'S1'</code>
4034		(default) or <code>'U1'</code> (if dtype=<code>'U'</code>)</p>
4035		</div>
4036
4037
4038		</section>
4039		<section id="stringtochar">
4040		<div class="attr function"><a class="headerlink" href="#stringtochar"># &nbsp</a>
4041
4042
4043		<span class="def">def</span>
4044		<span class="name">stringtochar</span><span class="signature">(unknown)</span>:
4045		</div>
4046
4047
4048		<div class="docstring"><p><strong><code>stringtochar(a,encoding='utf-8')</code></strong></p>
4049
4050		<p>convert a string array to a character array with one extra dimension</p>
4051
4052		<p><strong><code>a</code></strong>: Input numpy string array with numpy datatype <code>'SN'</code> or <code>'UN'</code>, where N
4053		is the number of characters in each string. Will be converted to
4054		an array of characters (datatype <code>'S1'</code> or <code>'U1'</code>) of shape <code>a.shape + (N,)</code>.</p>
4055
4056		<p>optional kwarg <code>encoding</code> can be used to specify character encoding (default
4057		<code>utf-8</code>). If <code>encoding</code> is 'none' or 'bytes', a <code>numpy.string_</code> the input array
4058		is treated a raw byte strings (<code>numpy.string_</code>).</p>
4059
4060		<p>returns a numpy character array with datatype <code>'S1'</code> or <code>'U1'</code>
4061		and shape <code>a.shape + (N,)</code>, where N is the length of each string in a.</p>
4062		</div>
4063
4064
4065		</section>
4066		<section id="chartostring">
4067		<div class="attr function"><a class="headerlink" href="#chartostring"># &nbsp</a>
4068
4069
4070		<span class="def">def</span>
4071		<span class="name">chartostring</span><span class="signature">(unknown)</span>:
4072		</div>
4073
4074
4075		<div class="docstring"><p><strong><code>chartostring(b,encoding='utf-8')</code></strong></p>
4076
4077		<p>convert a character array to a string array with one less dimension.</p>
4078
4079		<p><strong><code>b</code></strong>: Input character array (numpy datatype <code>'S1'</code> or <code>'U1'</code>).
4080		Will be converted to a array of strings, where each string has a fixed
4081		length of <code>b.shape[-1]</code> characters.</p>
4082
4083		<p>optional kwarg <code>encoding</code> can be used to specify character encoding (default
4084		<code>utf-8</code>). If <code>encoding</code> is 'none' or 'bytes', a <code>numpy.string_</code> btye array is
4085		returned.</p>
4086
4087		<p>returns a numpy string array with datatype <code>'UN'</code> (or <code>'SN'</code>) and shape
4088		<code>b.shape[:-1]</code> where where <code>N=b.shape[-1]</code>.</p>
4089		</div>
4090
4091
4092		</section>
4093	3754	<section id="MFDataset">
4094	3755	<div class="attr class">
4095	3756	<a class="headerlink" href="#MFDataset"># &nbsp</a>
4096	3757
4097	3758
4098	3759	<span class="def">class</span>
4099		<span class="name">MFDataset</span>(<span class="base"><a href="#Dataset">Dataset</a></span>):
	3760	<span class="name">MFDataset</span>(<span class="base"><a href="#Dataset">netCDF4.Dataset</a></span>):
4100	3761	</div>
4101	3762
4102	3763

4107	3768
4108	3769	<p>Adapted from <a href="http://pysclint.sourceforge.net/pycdf">pycdf</a> by Andre Gosselin.</p>
4109	3770
4110		<p>Example usage (See <a href="#<a href="#MFDataset.__init__">MFDataset.__init__</a>"><a href="#MFDataset.__init__">MFDataset.__init__</a></a> for more details):</p>
	3771	<p>Example usage (See <code><a href="#MFDataset.__init__">MFDataset.__init__</a></code> for more details):</p>
4111	3772
4112	3773	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
4113	3774	<span class="o">>>></span> <span class="c1"># create a series of netCDF files with a variable sharing</span>

4133	3794	<div class="attr function"><a class="headerlink" href="#MFDataset.__init__"># &nbsp</a>
4134	3795
4135	3796
4136		<span class="name">MFDataset</span><span class="signature">(files, check=False, aggdim=None, exclude=[], master_file=None)</span> </div>
	3797	<span class="name">MFDataset</span><span class="signature">(files, check=False, aggdim=None, exclude=[], master_file=None)</span>
	3798	</div>
4137	3799
4138	3800
4139	3801	<div class="docstring"><p><strong><code>__init__(self, files, check=False, aggdim=None, exclude=[],

4207	3869	<div class="inherited">
4208	3870	<h5>Inherited Members</h5>
4209	3871	<dl>
4210		<div><dt><a href="#Dataset">netCDF4._netCDF4.Dataset</a></dt>
	3872	<div><dt><a href="#Dataset">Dataset</a></dt>
4211	3873	<dd id="MFDataset.filepath" class="function"><a href="#Dataset.filepath">filepath</a></dd>
4212	3874	<dd id="MFDataset.isopen" class="function"><a href="#Dataset.isopen">isopen</a></dd>
4213	3875	<dd id="MFDataset.sync" class="function"><a href="#Dataset.sync">sync</a></dd>

4261	3923
4262	3924
4263	3925	<span class="def">class</span>
4264		<span class="name">MFTime</span>(<span class="base"><a href="#_Variable">_Variable</a></span>):
	3926	<span class="name">MFTime</span>(<span class="base">netCDF4._netCDF4._Variable</span>):
4265	3927	</div>
4266	3928
4267	3929
4268	3930	<div class="docstring"><p>Class providing an interface to a MFDataset time Variable by imposing a unique common
4269	3931	time unit and/or calendar to all files.</p>
4270	3932
4271		<p>Example usage (See <a href="#<a href="#MFTime.__init__">MFTime.__init__</a>"><a href="#MFTime.__init__">MFTime.__init__</a></a> for more details):</p>
4272
4273		<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="kn">import</span> <span class="nn">numpy</span>
	3933	<p>Example usage (See <code><a href="#MFTime.__init__">MFTime.__init__</a></code> for more details):</p>
	3934
	3935	<div class="codehilite"><pre><span></span><code><span class="o">>>></span> <span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
4274	3936	<span class="o">>>></span> <span class="n">f1</span> <span class="o">=</span> <span class="n">Dataset</span><span class="p">(</span><span class="s2">"mftest_1.nc"</span><span class="p">,</span><span class="s2">"w"</span><span class="p">,</span> <span class="nb">format</span><span class="o">=</span><span class="s2">"NETCDF4_CLASSIC"</span><span class="p">)</span>
4275	3937	<span class="o">>>></span> <span class="n">f2</span> <span class="o">=</span> <span class="n">Dataset</span><span class="p">(</span><span class="s2">"mftest_2.nc"</span><span class="p">,</span><span class="s2">"w"</span><span class="p">,</span> <span class="nb">format</span><span class="o">=</span><span class="s2">"NETCDF4_CLASSIC"</span><span class="p">)</span>
4276	3938	<span class="o">>>></span> <span class="n">f1</span><span class="o">.</span><span class="n">createDimension</span><span class="p">(</span><span class="s2">"time"</span><span class="p">,</span><span class="kc">None</span><span class="p">)</span>

4281	3943	<span class="o">>>></span> <span class="n">t2</span><span class="o">.</span><span class="n">units</span> <span class="o">=</span> <span class="s2">"days since 2000-02-01"</span>
4282	3944	<span class="o">>>></span> <span class="n">t1</span><span class="o">.</span><span class="n">calendar</span> <span class="o">=</span> <span class="s2">"standard"</span>
4283	3945	<span class="o">>>></span> <span class="n">t2</span><span class="o">.</span><span class="n">calendar</span> <span class="o">=</span> <span class="s2">"standard"</span>
4284		<span class="o">>>></span> <span class="n">t1</span><span class="p">[:]</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">31</span><span class="p">)</span>
4285		<span class="o">>>></span> <span class="n">t2</span><span class="p">[:]</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">30</span><span class="p">)</span>
	3946	<span class="o">>>></span> <span class="n">t1</span><span class="p">[:]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">31</span><span class="p">)</span>
	3947	<span class="o">>>></span> <span class="n">t2</span><span class="p">[:]</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">arange</span><span class="p">(</span><span class="mi">30</span><span class="p">)</span>
4286	3948	<span class="o">>>></span> <span class="n">f1</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
4287	3949	<span class="o">>>></span> <span class="n">f2</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
4288	3950	<span class="o">>>></span> <span class="c1"># Read the two files in at once, in one Dataset.</span>

4303	3965	<div class="attr function"><a class="headerlink" href="#MFTime.__init__"># &nbsp</a>
4304	3966
4305	3967
4306		<span class="name">MFTime</span><span class="signature">(time, units=None, calendar=None)</span> </div>
	3968	<span class="name">MFTime</span><span class="signature">(time, units=None, calendar=None)</span>
	3969	</div>
4307	3970
4308	3971
4309	3972	<div class="docstring"><p><strong><code>__init__(self, time, units=None, calendar=None)</code></strong></p>

4311	3974	<p>Create a time Variable with units consistent across a multifile
4312	3975	dataset.</p>
4313	3976
4314		<p><strong><code>time</code></strong>: Time variable from a <a href="#MFDataset">MFDataset</a>.</p>
	3977	<p><strong><code>time</code></strong>: Time variable from a <code><a href="#MFDataset">MFDataset</a></code>.</p>
4315	3978
4316	3979	<p><strong><code>units</code></strong>: Time units, for example, <code>'days since 1979-01-01'</code>. If <code>None</code>,
4317	3980	use the units from the master variable.</p>

4327	3990	<div class="inherited">
4328	3991	<h5>Inherited Members</h5>
4329	3992	<dl>
4330		<div><dt><a href="#_Variable">netCDF4._netCDF4._Variable</a></dt>
4331		<dd id="MFTime.typecode" class="function"><a href="#_Variable.typecode">typecode</a></dd>
4332		<dd id="MFTime.ncattrs" class="function"><a href="#_Variable.ncattrs">ncattrs</a></dd>
4333		<dd id="MFTime.set_auto_chartostring" class="function"><a href="#_Variable.set_auto_chartostring">set_auto_chartostring</a></dd>
4334		<dd id="MFTime.set_auto_maskandscale" class="function"><a href="#_Variable.set_auto_maskandscale">set_auto_maskandscale</a></dd>
4335		<dd id="MFTime.set_auto_mask" class="function"><a href="#_Variable.set_auto_mask">set_auto_mask</a></dd>
4336		<dd id="MFTime.set_auto_scale" class="function"><a href="#_Variable.set_auto_scale">set_auto_scale</a></dd>
4337		<dd id="MFTime.set_always_mask" class="function"><a href="#_Variable.set_always_mask">set_always_mask</a></dd>
	3993	<div><dt>netCDF4._netCDF4._Variable</dt>
	3994	<dd id="MFTime.typecode" class="function">typecode</dd>
	3995	<dd id="MFTime.ncattrs" class="function">ncattrs</dd>
	3996	<dd id="MFTime.set_auto_chartostring" class="function">set_auto_chartostring</dd>
	3997	<dd id="MFTime.set_auto_maskandscale" class="function">set_auto_maskandscale</dd>
	3998	<dd id="MFTime.set_auto_mask" class="function">set_auto_mask</dd>
	3999	<dd id="MFTime.set_auto_scale" class="function">set_auto_scale</dd>
	4000	<dd id="MFTime.set_always_mask" class="function">set_always_mask</dd>
4338	4001
4339	4002	</div>
4340	4003	</dl>
4341	4004	</div>
	4005	</section>
	4006	<section id="CompoundType">
	4007	<div class="attr class">
	4008	<a class="headerlink" href="#CompoundType"># &nbsp</a>
	4009
	4010
	4011	<span class="def">class</span>
	4012	<span class="name">CompoundType</span>:
	4013	</div>
	4014
	4015
	4016	<div class="docstring"><p>A <code><a href="#CompoundType">CompoundType</a></code> instance is used to describe a compound data
	4017	type, and can be passed to the the <code><a href="#Dataset.createVariable">Dataset.createVariable</a></code> method of
	4018	a <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> instance.
	4019	Compound data types map to numpy structured arrays.
	4020	See <code><a href="#CompoundType.__init__">CompoundType.__init__</a></code> for more details.</p>
	4021
	4022	<p>The instance variables <code><a href="#CompoundType.dtype">dtype</a></code> and <code><a href="#CompoundType.name">name</a></code> should not be modified by
	4023	the user.</p>
	4024	</div>
	4025
	4026
	4027	<div id="CompoundType.__init__" class="classattr">
	4028	<div class="attr function"><a class="headerlink" href="#CompoundType.__init__"># &nbsp</a>
	4029
	4030
	4031	<span class="name">CompoundType</span><span class="signature">()</span>
	4032	</div>
	4033
	4034
	4035	<div class="docstring"><p><strong><em><code>__init__(group, datatype, datatype_name)</code></em></strong></p>
	4036
	4037	<p>CompoundType constructor.</p>
	4038
	4039	<p><strong><code>group</code></strong>: <code><a href="#Group">Group</a></code> instance to associate with the compound datatype.</p>
	4040
	4041	<p><strong><code>datatype</code></strong>: A numpy dtype object describing a structured (a.k.a record)
	4042	array. Can be composed of homogeneous numeric or character data types, or
	4043	other structured array data types.</p>
	4044
	4045	<p><strong><code>datatype_name</code></strong>: a Python string containing a description of the
	4046	compound data type.</p>
	4047
	4048	<p><strong><em>Note 1</em></strong>: When creating nested compound data types,
	4049	the inner compound data types must already be associated with CompoundType
	4050	instances (so create CompoundType instances for the innermost structures
	4051	first).</p>
	4052
	4053	<p><strong><em>Note 2</em></strong>: <code><a href="#CompoundType">CompoundType</a></code> instances should be created using the
	4054	<code><a href="#Dataset.createCompoundType">Dataset.createCompoundType</a></code> method of a <code><a href="#Dataset">Dataset</a></code> or
	4055	<code><a href="#Group">Group</a></code> instance, not using this class directly.</p>
	4056	</div>
	4057
	4058
	4059	</div>
	4060	<div id="CompoundType.dtype" class="classattr">
	4061	<div class="attr variable"><a class="headerlink" href="#CompoundType.dtype"># &nbsp</a>
	4062
	4063	<span class="name">dtype</span><span class="default_value"> = <attribute 'dtype' of '<a href="#_netCDF4.CompoundType">netCDF4._netCDF4.CompoundType</a>' objects></span>
	4064	</div>
	4065
	4066
	4067
	4068	</div>
	4069	<div id="CompoundType.dtype_view" class="classattr">
	4070	<div class="attr variable"><a class="headerlink" href="#CompoundType.dtype_view"># &nbsp</a>
	4071
	4072	<span class="name">dtype_view</span><span class="default_value"> = <attribute 'dtype_view' of '<a href="#_netCDF4.CompoundType">netCDF4._netCDF4.CompoundType</a>' objects></span>
	4073	</div>
	4074
	4075
	4076
	4077	</div>
	4078	<div id="CompoundType.name" class="classattr">
	4079	<div class="attr variable"><a class="headerlink" href="#CompoundType.name"># &nbsp</a>
	4080
	4081	<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#_netCDF4.CompoundType">netCDF4._netCDF4.CompoundType</a>' objects></span>
	4082	</div>
	4083
	4084
	4085
	4086	</div>
	4087	</section>
	4088	<section id="VLType">
	4089	<div class="attr class">
	4090	<a class="headerlink" href="#VLType"># &nbsp</a>
	4091
	4092
	4093	<span class="def">class</span>
	4094	<span class="name">VLType</span>:
	4095	</div>
	4096
	4097
	4098	<div class="docstring"><p>A <code><a href="#VLType">VLType</a></code> instance is used to describe a variable length (VLEN) data
	4099	type, and can be passed to the the <code><a href="#Dataset.createVariable">Dataset.createVariable</a></code> method of
	4100	a <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> instance. See
	4101	<code><a href="#VLType.__init__">VLType.__init__</a></code> for more details.</p>
	4102
	4103	<p>The instance variables <code><a href="#VLType.dtype">dtype</a></code> and <code><a href="#VLType.name">name</a></code> should not be modified by
	4104	the user.</p>
	4105	</div>
	4106
	4107
	4108	<div id="VLType.__init__" class="classattr">
	4109	<div class="attr function"><a class="headerlink" href="#VLType.__init__"># &nbsp</a>
	4110
	4111
	4112	<span class="name">VLType</span><span class="signature">()</span>
	4113	</div>
	4114
	4115
	4116	<div class="docstring"><p><strong><code>__init__(group, datatype, datatype_name)</code></strong></p>
	4117
	4118	<p>VLType constructor.</p>
	4119
	4120	<p><strong><code>group</code></strong>: <code><a href="#Group">Group</a></code> instance to associate with the VLEN datatype.</p>
	4121
	4122	<p><strong><code>datatype</code></strong>: An numpy dtype object describing the component type for the
	4123	variable length array.</p>
	4124
	4125	<p><strong><code>datatype_name</code></strong>: a Python string containing a description of the
	4126	VLEN data type.</p>
	4127
	4128	<p><strong><em><code>Note</code></em></strong>: <code><a href="#VLType">VLType</a></code> instances should be created using the
	4129	<code><a href="#Dataset.createVLType">Dataset.createVLType</a></code> method of a <code><a href="#Dataset">Dataset</a></code> or
	4130	<code><a href="#Group">Group</a></code> instance, not using this class directly.</p>
	4131	</div>
	4132
	4133
	4134	</div>
	4135	<div id="VLType.dtype" class="classattr">
	4136	<div class="attr variable"><a class="headerlink" href="#VLType.dtype"># &nbsp</a>
	4137
	4138	<span class="name">dtype</span><span class="default_value"> = <attribute 'dtype' of '<a href="#_netCDF4.VLType">netCDF4._netCDF4.VLType</a>' objects></span>
	4139	</div>
	4140
	4141
	4142
	4143	</div>
	4144	<div id="VLType.name" class="classattr">
	4145	<div class="attr variable"><a class="headerlink" href="#VLType.name"># &nbsp</a>
	4146
	4147	<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#_netCDF4.VLType">netCDF4._netCDF4.VLType</a>' objects></span>
	4148	</div>
	4149
	4150
	4151
	4152	</div>
	4153	</section>
	4154	<section id="date2num">
	4155	<div class="attr function"><a class="headerlink" href="#date2num"># &nbsp</a>
	4156
	4157
	4158	<span class="def">def</span>
	4159	<span class="name">date2num</span><span class="signature">(unknown)</span>:
	4160	</div>
	4161
	4162
	4163	<div class="docstring"><p>date2num(dates, units, calendar=None, has_year_zero=None)</p>
	4164
	4165	<p>Return numeric time values given datetime objects. The units
	4166	of the numeric time values are described by the <strong>units</strong> argument
	4167	and the <strong>calendar</strong> keyword. The datetime objects must
	4168	be in UTC with no time-zone offset. If there is a
	4169	time-zone offset in <strong>units</strong>, it will be applied to the
	4170	returned numeric values.</p>
	4171
	4172	<p><strong>dates</strong>: A datetime object or a sequence of datetime objects.
	4173	The datetime objects should not include a time-zone offset. They
	4174	can be either native python datetime instances (which use
	4175	the proleptic gregorian calendar) or cftime.datetime instances.</p>
	4176
	4177	<p><strong>units</strong>: a string of the form <strong><time units> since <reference time></strong>
	4178	describing the time units. <strong><time units></strong> can be days, hours, minutes,
	4179	seconds, milliseconds or microseconds. <strong><reference time></strong> is the time
	4180	origin. <strong>months_since</strong> is allowed <em>only</em> for the <strong>360_day</strong> calendar.</p>
	4181
	4182	<p><strong>calendar</strong>: describes the calendar to be used in the time calculations.
	4183	All the values currently defined in the
	4184	<code>CF metadata convention <http://cfconventions.org/cf-conventions/cf-conventions#calendar></code>__ are supported.
	4185	Valid calendars <strong>'standard', 'gregorian', 'proleptic_gregorian'
	4186	'noleap', '365_day', '360_day', 'julian', 'all_leap', '366_day'</strong>.
	4187	Default is <code>None</code> which means the calendar associated with the first
	4188	input datetime instance will be used.</p>
	4189
	4190	<p><strong>has_year_zero</strong>: If set to True, astronomical year numbering
	4191	is used and the year zero exists. If set to False for real-world
	4192	calendars, then historical year numbering is used and the year 1 is
	4193	preceded by year -1 and no year zero exists.
	4194	The defaults are set to conform with
	4195	CF version 1.9 conventions (False for 'julian', 'gregorian'/'standard', True
	4196	for 'proleptic_gregorian' (ISO 8601) and True for the idealized
	4197	calendars 'noleap'/'365_day', '360_day', 366_day'/'all_leap')
	4198	The defaults can only be over-ridden for the real-world calendars,
	4199	for the the idealized calendars the year zero
	4200	always exists and the has_year_zero kwarg is ignored.
	4201	This kwarg is not needed to define calendar systems allowed by CF
	4202	(the calendar-specific defaults do this).</p>
	4203
	4204	<p>returns a numeric time value, or an array of numeric time values
	4205	with approximately 1 microsecond accuracy.</p>
	4206	</div>
	4207
	4208
	4209	</section>
	4210	<section id="num2date">
	4211	<div class="attr function"><a class="headerlink" href="#num2date"># &nbsp</a>
	4212
	4213
	4214	<span class="def">def</span>
	4215	<span class="name">num2date</span><span class="signature">(unknown)</span>:
	4216	</div>
	4217
	4218
	4219	<div class="docstring"><p>num2date(times, units, calendar=u'standard', only_use_cftime_datetimes=True, only_use_python_datetimes=False, has_year_zero=None)</p>
	4220
	4221	<p>Return datetime objects given numeric time values. The units
	4222	of the numeric time values are described by the <strong>units</strong> argument
	4223	and the <strong>calendar</strong> keyword. The returned datetime objects represent
	4224	UTC with no time-zone offset, even if the specified
	4225	<strong>units</strong> contain a time-zone offset.</p>
	4226
	4227	<p><strong>times</strong>: numeric time values.</p>
	4228
	4229	<p><strong>units</strong>: a string of the form <strong><time units> since <reference time></strong>
	4230	describing the time units. <strong><time units></strong> can be days, hours, minutes,
	4231	seconds, milliseconds or microseconds. <strong><reference time></strong> is the time
	4232	origin. <strong>months_since</strong> is allowed <em>only</em> for the <strong>360_day</strong> calendar.</p>
	4233
	4234	<p><strong>calendar</strong>: describes the calendar used in the time calculations.
	4235	All the values currently defined in the
	4236	<code>CF metadata convention <http://cfconventions.org/cf-conventions/cf-conventions#calendar></code>__ are supported.
	4237	Valid calendars <strong>'standard', 'gregorian', 'proleptic_gregorian'
	4238	'noleap', '365_day', '360_day', 'julian', 'all_leap', '366_day'</strong>.
	4239	Default is <strong>'standard'</strong>, which is a mixed Julian/Gregorian calendar.</p>
	4240
	4241	<p><strong>only_use_cftime_datetimes</strong>: if False, python datetime.datetime
	4242	objects are returned from num2date where possible; if True dates which
	4243	subclass cftime.datetime are returned for all calendars. Default <strong>True</strong>.</p>
	4244
	4245	<p><strong>only_use_python_datetimes</strong>: always return python datetime.datetime
	4246	objects and raise an error if this is not possible. Ignored unless
	4247	<strong>only_use_cftime_datetimes=False</strong>. Default <strong>False</strong>.</p>
	4248
	4249	<p><strong>has_year_zero</strong>: if set to True, astronomical year numbering
	4250	is used and the year zero exists. If set to False for real-world
	4251	calendars, then historical year numbering is used and the year 1 is
	4252	preceded by year -1 and no year zero exists.
	4253	The defaults are set to conform with
	4254	CF version 1.9 conventions (False for 'julian', 'gregorian'/'standard', True
	4255	for 'proleptic_gregorian' (ISO 8601) and True for the idealized
	4256	calendars 'noleap'/'365_day', '360_day', 366_day'/'all_leap')
	4257	The defaults can only be over-ridden for the real-world calendars,
	4258	for the the idealized calendars the year zero
	4259	always exists and the has_year_zero kwarg is ignored.
	4260	This kwarg is not needed to define calendar systems allowed by CF
	4261	(the calendar-specific defaults do this).</p>
	4262
	4263	<p>returns a datetime instance, or an array of datetime instances with
	4264	microsecond accuracy, if possible.</p>
	4265
	4266	<p><strong><em>Note</em></strong>: If only_use_cftime_datetimes=False and
	4267	use_only_python_datetimes=False, the datetime instances
	4268	returned are 'real' python datetime
	4269	objects if <strong>calendar='proleptic_gregorian'</strong>, or
	4270	<strong>calendar='standard'</strong> or <strong>'gregorian'</strong>
	4271	and the date is after the breakpoint between the Julian and
	4272	Gregorian calendars (1582-10-15). Otherwise, they are ctime.datetime
	4273	objects which support some but not all the methods of native python
	4274	datetime objects. The datetime instances
	4275	do not contain a time-zone offset, even if the specified <strong>units</strong>
	4276	contains one.</p>
	4277	</div>
	4278
	4279
	4280	</section>
	4281	<section id="date2index">
	4282	<div class="attr function"><a class="headerlink" href="#date2index"># &nbsp</a>
	4283
	4284
	4285	<span class="def">def</span>
	4286	<span class="name">date2index</span><span class="signature">(unknown)</span>:
	4287	</div>
	4288
	4289
	4290	<div class="docstring"><p>date2index(dates, nctime, calendar=None, select=u'exact', has_year_zero=None)</p>
	4291
	4292	<p>Return indices of a netCDF time variable corresponding to the given dates.</p>
	4293
	4294	<p><strong>dates</strong>: A datetime object or a sequence of datetime objects.
	4295	The datetime objects should not include a time-zone offset.</p>
	4296
	4297	<p><strong>nctime</strong>: A netCDF time variable object. The nctime object must have a
	4298	<strong>units</strong> attribute.</p>
	4299
	4300	<p><strong>calendar</strong>: describes the calendar to be used in the time calculations.
	4301	All the values currently defined in the
	4302	<code>CF metadata convention <http://cfconventions.org/cf-conventions/cf-conventions#calendar></code>__ are supported.
	4303	Valid calendars <strong>'standard', 'gregorian', 'proleptic_gregorian'
	4304	'noleap', '365_day', '360_day', 'julian', 'all_leap', '366_day'</strong>.
	4305	Default is <code>None</code> which means the calendar associated with the first
	4306	input datetime instance will be used.</p>
	4307
	4308	<p><strong>select</strong>: <strong>'exact', 'before', 'after', 'nearest'</strong>
	4309	The index selection method. <strong>exact</strong> will return the indices perfectly
	4310	matching the dates given. <strong>before</strong> and <strong>after</strong> will return the indices
	4311	corresponding to the dates just before or just after the given dates if
	4312	an exact match cannot be found. <strong>nearest</strong> will return the indices that
	4313	correspond to the closest dates.</p>
	4314
	4315	<p><strong>has_year_zero</strong>: if set to True, astronomical year numbering
	4316	is used and the year zero exists. If set to False for real-world
	4317	calendars, then historical year numbering is used and the year 1 is
	4318	preceded by year -1 and no year zero exists.
	4319	The defaults are set to conform with
	4320	CF version 1.9 conventions (False for 'julian', 'gregorian'/'standard', True
	4321	for 'proleptic_gregorian' (ISO 8601) and True for the idealized
	4322	calendars 'noleap'/'365_day', '360_day', 366_day'/'all_leap')
	4323	The defaults can only be over-ridden for the real-world calendars,
	4324	for the the idealized calendars the year zero
	4325	always exists and the has_year_zero kwarg is ignored.
	4326	This kwarg is not needed to define calendar systems allowed by CF
	4327	(the calendar-specific defaults do this).</p>
	4328
	4329	<p>returns an index (indices) of the netCDF time variable corresponding
	4330	to the given datetime object(s).</p>
	4331	</div>
	4332
	4333
	4334	</section>
	4335	<section id="stringtochar">
	4336	<div class="attr function"><a class="headerlink" href="#stringtochar"># &nbsp</a>
	4337
	4338
	4339	<span class="def">def</span>
	4340	<span class="name">stringtochar</span><span class="signature">(unknown)</span>:
	4341	</div>
	4342
	4343
	4344	<div class="docstring"><p><strong><code>stringtochar(a,encoding='utf-8')</code></strong></p>
	4345
	4346	<p>convert a string array to a character array with one extra dimension</p>
	4347
	4348	<p><strong><code>a</code></strong>: Input numpy string array with numpy datatype <code>'SN'</code> or <code>'UN'</code>, where N
	4349	is the number of characters in each string. Will be converted to
	4350	an array of characters (datatype <code>'S1'</code> or <code>'U1'</code>) of shape <code>a.shape + (N,)</code>.</p>
	4351
	4352	<p>optional kwarg <code>encoding</code> can be used to specify character encoding (default
	4353	<code>utf-8</code>). If <code>encoding</code> is 'none' or 'bytes', a <code>numpy.string_</code> the input array
	4354	is treated a raw byte strings (<code>numpy.string_</code>).</p>
	4355
	4356	<p>returns a numpy character array with datatype <code>'S1'</code> or <code>'U1'</code>
	4357	and shape <code>a.shape + (N,)</code>, where N is the length of each string in a.</p>
	4358	</div>
	4359
	4360
	4361	</section>
	4362	<section id="chartostring">
	4363	<div class="attr function"><a class="headerlink" href="#chartostring"># &nbsp</a>
	4364
	4365
	4366	<span class="def">def</span>
	4367	<span class="name">chartostring</span><span class="signature">(unknown)</span>:
	4368	</div>
	4369
	4370
	4371	<div class="docstring"><p><strong><code>chartostring(b,encoding='utf-8')</code></strong></p>
	4372
	4373	<p>convert a character array to a string array with one less dimension.</p>
	4374
	4375	<p><strong><code>b</code></strong>: Input character array (numpy datatype <code>'S1'</code> or <code>'U1'</code>).
	4376	Will be converted to a array of strings, where each string has a fixed
	4377	length of <code>b.shape[-1]</code> characters.</p>
	4378
	4379	<p>optional kwarg <code>encoding</code> can be used to specify character encoding (default
	4380	<code>utf-8</code>). If <code>encoding</code> is 'none' or 'bytes', a <code>numpy.string_</code> btye array is
	4381	returned.</p>
	4382
	4383	<p>returns a numpy string array with datatype <code>'UN'</code> (or <code>'SN'</code>) and shape
	4384	<code>b.shape[:-1]</code> where where <code>N=b.shape[-1]</code>.</p>
	4385	</div>
	4386
	4387
	4388	</section>
	4389	<section id="stringtoarr">
	4390	<div class="attr function"><a class="headerlink" href="#stringtoarr"># &nbsp</a>
	4391
	4392
	4393	<span class="def">def</span>
	4394	<span class="name">stringtoarr</span><span class="signature">(unknown)</span>:
	4395	</div>
	4396
	4397
	4398	<div class="docstring"><p><strong><code>stringtoarr(a, NUMCHARS,dtype='S')</code></strong></p>
	4399
	4400	<p>convert a string to a character array of length <code>NUMCHARS</code></p>
	4401
	4402	<p><strong><code>a</code></strong>: Input python string.</p>
	4403
	4404	<p><strong><code>NUMCHARS</code></strong>: number of characters used to represent string
	4405	(if len(a) < <code>NUMCHARS</code>, it will be padded on the right with blanks).</p>
	4406
	4407	<p><strong><code>dtype</code></strong>: type of numpy array to return. Default is <code>'S'</code>, which
	4408	means an array of dtype <code>'S1'</code> will be returned. If dtype=<code>'U'</code>, a
	4409	unicode array (dtype = <code>'U1'</code>) will be returned.</p>
	4410
	4411	<p>returns a rank 1 numpy character array of length NUMCHARS with datatype <code>'S1'</code>
	4412	(default) or <code>'U1'</code> (if dtype=<code>'U'</code>)</p>
	4413	</div>
	4414
	4415
	4416	</section>
	4417	<section id="getlibversion">
	4418	<div class="attr function"><a class="headerlink" href="#getlibversion"># &nbsp</a>
	4419
	4420
	4421	<span class="def">def</span>
	4422	<span class="name">getlibversion</span><span class="signature">(unknown)</span>:
	4423	</div>
	4424
	4425
	4426	<div class="docstring"><p><strong><code><a href="#getlibversion">getlibversion()</a></code></strong></p>
	4427
	4428	<p>returns a string describing the version of the netcdf library
	4429	used to build the module, and when it was built.</p>
	4430	</div>
	4431
	4432
	4433	</section>
	4434	<section id="EnumType">
	4435	<div class="attr class">
	4436	<a class="headerlink" href="#EnumType"># &nbsp</a>
	4437
	4438
	4439	<span class="def">class</span>
	4440	<span class="name">EnumType</span>:
	4441	</div>
	4442
	4443
	4444	<div class="docstring"><p>A <code><a href="#EnumType">EnumType</a></code> instance is used to describe an Enum data
	4445	type, and can be passed to the the <code><a href="#Dataset.createVariable">Dataset.createVariable</a></code> method of
	4446	a <code><a href="#Dataset">Dataset</a></code> or <code><a href="#Group">Group</a></code> instance. See
	4447	<code><a href="#EnumType.__init__">EnumType.__init__</a></code> for more details.</p>
	4448
	4449	<p>The instance variables <code><a href="#EnumType.dtype">dtype</a></code>, <code><a href="#EnumType.name">name</a></code> and <code><a href="#EnumType.enum_dict">enum_dict</a></code> should not be modified by
	4450	the user.</p>
	4451	</div>
	4452
	4453
	4454	<div id="EnumType.__init__" class="classattr">
	4455	<div class="attr function"><a class="headerlink" href="#EnumType.__init__"># &nbsp</a>
	4456
	4457
	4458	<span class="name">EnumType</span><span class="signature">()</span>
	4459	</div>
	4460
	4461
	4462	<div class="docstring"><p><strong><code>__init__(group, datatype, datatype_name, enum_dict)</code></strong></p>
	4463
	4464	<p>EnumType constructor.</p>
	4465
	4466	<p><strong><code>group</code></strong>: <code><a href="#Group">Group</a></code> instance to associate with the VLEN datatype.</p>
	4467
	4468	<p><strong><code>datatype</code></strong>: An numpy integer dtype object describing the base type
	4469	for the Enum.</p>
	4470
	4471	<p><strong><code>datatype_name</code></strong>: a Python string containing a description of the
	4472	Enum data type.</p>
	4473
	4474	<p><strong><code><a href="#EnumType.enum_dict">enum_dict</a></code></strong>: a Python dictionary containing the Enum field/value
	4475	pairs.</p>
	4476
	4477	<p><strong><em><code>Note</code></em></strong>: <code><a href="#EnumType">EnumType</a></code> instances should be created using the
	4478	<code><a href="#Dataset.createEnumType">Dataset.createEnumType</a></code> method of a <code><a href="#Dataset">Dataset</a></code> or
	4479	<code><a href="#Group">Group</a></code> instance, not using this class directly.</p>
	4480	</div>
	4481
	4482
	4483	</div>
	4484	<div id="EnumType.dtype" class="classattr">
	4485	<div class="attr variable"><a class="headerlink" href="#EnumType.dtype"># &nbsp</a>
	4486
	4487	<span class="name">dtype</span><span class="default_value"> = <attribute 'dtype' of '<a href="#_netCDF4.EnumType">netCDF4._netCDF4.EnumType</a>' objects></span>
	4488	</div>
	4489
	4490
	4491
	4492	</div>
	4493	<div id="EnumType.name" class="classattr">
	4494	<div class="attr variable"><a class="headerlink" href="#EnumType.name"># &nbsp</a>
	4495
	4496	<span class="name">name</span><span class="default_value"> = <attribute 'name' of '<a href="#_netCDF4.EnumType">netCDF4._netCDF4.EnumType</a>' objects></span>
	4497	</div>
	4498
	4499
	4500
	4501	</div>
	4502	<div id="EnumType.enum_dict" class="classattr">
	4503	<div class="attr variable"><a class="headerlink" href="#EnumType.enum_dict"># &nbsp</a>
	4504
	4505	<span class="name">enum_dict</span><span class="default_value"> = <attribute 'enum_dict' of '<a href="#_netCDF4.EnumType">netCDF4._netCDF4.EnumType</a>' objects></span>
	4506	</div>
	4507
	4508
	4509
	4510	</div>
	4511	</section>
	4512	<section id="get_chunk_cache">
	4513	<div class="attr function"><a class="headerlink" href="#get_chunk_cache"># &nbsp</a>
	4514
	4515
	4516	<span class="def">def</span>
	4517	<span class="name">get_chunk_cache</span><span class="signature">(unknown)</span>:
	4518	</div>
	4519
	4520
	4521	<div class="docstring"><p><strong><code><a href="#get_chunk_cache">get_chunk_cache()</a></code></strong></p>
	4522
	4523	<p>return current netCDF chunk cache information in a tuple (size,nelems,preemption).
	4524	See netcdf C library documentation for <code>nc_get_chunk_cache</code> for
	4525	details. Values can be reset with <code><a href="#set_chunk_cache">set_chunk_cache</a></code>.</p>
	4526	</div>
	4527
	4528
	4529	</section>
	4530	<section id="set_chunk_cache">
	4531	<div class="attr function"><a class="headerlink" href="#set_chunk_cache"># &nbsp</a>
	4532
	4533
	4534	<span class="def">def</span>
	4535	<span class="name">set_chunk_cache</span><span class="signature">(unknown)</span>:
	4536	</div>
	4537
	4538
	4539	<div class="docstring"><p><strong><code>set_chunk_cache(self,size=None,nelems=None,preemption=None)</code></strong></p>
	4540
	4541	<p>change netCDF4 chunk cache settings.
	4542	See netcdf C library documentation for <code>nc_set_chunk_cache</code> for
	4543	details.</p>
	4544	</div>
	4545
	4546
4342	4547	</section>
4343	4548	</main>
4344	4549	</body>

+18

-0

examples/mpi_example_compressed.py less more

	0	# to run: mpirun -np 4 python mpi_example_compressed.py
	1	import sys
	2	from mpi4py import MPI
	3	import numpy as np
	4	from netCDF4 import Dataset
	5	rank = MPI.COMM_WORLD.rank # The process ID (integer 0-3 for 4-process run)
	6	nc = Dataset('parallel_test_compressed.nc', 'w', parallel=True)
	7	d = nc.createDimension('dim',4)
	8	v = nc.createVariable('var', np.int32, 'dim', zlib=True)
	9	v[:] = np.arange(4)
	10	nc.close()
	11	# read compressed files in parallel, check the data, try to rewrite some data
	12	nc = Dataset('parallel_test_compressed.nc', 'a', parallel=True)
	13	v = nc['var']
	14	assert rank==v[rank]
	15	v.set_collective(True) # issue #1108 (var must be in collective mode or write will fail)
	16	v[rank]=2*rank
	17	nc.close()

-1293

~~examples/reading_netCDF.ipynb~~ less more

0		{
1		"cells": [
2		{
3		"cell_type": "markdown",
4		"metadata": {
5		"internals": {
6		"slide_helper": "subslide_end",
7		"slide_type": "subslide"
8		},
9		"slide_helper": "slide_end",
10		"slideshow": {
11		"slide_type": "slide"
12		}
13		},
14		"source": [
15		"# Reading netCDF data\n",
16		"- requires [numpy](http://numpy.scipy.org) and netCDF/HDF5 C libraries.\n",
17		"- Github site: https://github.com/Unidata/netcdf4-python\n",
18		"- Online docs: http://unidata.github.io/netcdf4-python/\n",
19		"- Based on Konrad Hinsen's old [Scientific.IO.NetCDF](http://dirac.cnrs-orleans.fr/plone/software/scientificpython/) API, with lots of added netcdf version 4 features.\n",
20		"- Developed by Jeff Whitaker at NOAA, with many contributions from users."
21		]
22		},
23		{
24		"cell_type": "markdown",
25		"metadata": {
26		"internals": {
27		"slide_type": "subslide"
28		},
29		"slideshow": {
30		"slide_type": "slide"
31		}
32		},
33		"source": [
34		"## Interactively exploring a netCDF File\n",
35		"\n",
36		"Let's explore a netCDF file from the Atlantic Real-Time Ocean Forecast System\n",
37		"\n",
38		"first, import netcdf4-python and numpy"
39		]
40		},
41		{
42		"cell_type": "code",
43		"execution_count": 5,
44		"metadata": {
45		"collapsed": false,
46		"internals": {
47		"frag_number": 2,
48		"slide_helper": "subslide_end"
49		},
50		"slide_helper": "slide_end",
51		"slideshow": {
52		"slide_type": "fragment"
53		}
54		},
55		"outputs": [],
56		"source": [
57		"import netCDF4\n",
58		"import numpy as np"
59		]
60		},
61		{
62		"cell_type": "markdown",
63		"metadata": {
64		"internals": {
65		"frag_helper": "fragment_end",
66		"frag_number": 2,
67		"slide_type": "subslide"
68		},
69		"slideshow": {
70		"slide_type": "slide"
71		}
72		},
73		"source": [
74		"## Create a netCDF4.Dataset object\n",
75		"- `f` is a `Dataset` object, representing an open netCDF file.\n",
76		"- printing the object gives you summary information, similar to `ncdump -h`."
77		]
78		},
79		{
80		"cell_type": "code",
81		"execution_count": 6,
82		"metadata": {
83		"collapsed": false,
84		"internals": {
85		"frag_helper": "fragment_end",
86		"frag_number": 4,
87		"slide_helper": "subslide_end"
88		},
89		"slide_helper": "slide_end",
90		"slideshow": {
91		"slide_type": "fragment"
92		}
93		},
94		"outputs": [
95		{
96		"name": "stdout",
97		"output_type": "stream",
98		"text": [
99		"<type 'netCDF4._netCDF4.Dataset'>\n",
100		"root group (NETCDF4_CLASSIC data model, file format HDF5):\n",
101		" Conventions: CF-1.0\n",
102		" title: HYCOM ATLb2.00\n",
103		" institution: National Centers for Environmental Prediction\n",
104		" source: HYCOM archive file\n",
105		" experiment: 90.9\n",
106		" history: archv2ncdf3z\n",
107		" dimensions(sizes): MT(1), Y(850), X(712), Depth(10)\n",
108		" variables(dimensions): float64 \u001b[4mMT\u001b[0m(MT), float64 \u001b[4mDate\u001b[0m(MT), float32 \u001b[4mDepth\u001b[0m(Depth), int32 \u001b[4mY\u001b[0m(Y), int32 \u001b[4mX\u001b[0m(X), float32 \u001b[4mLatitude\u001b[0m(Y,X), float32 \u001b[4mLongitude\u001b[0m(Y,X), float32 \u001b[4mu\u001b[0m(MT,Depth,Y,X), float32 \u001b[4mv\u001b[0m(MT,Depth,Y,X), float32 \u001b[4mtemperature\u001b[0m(MT,Depth,Y,X), float32 \u001b[4msalinity\u001b[0m(MT,Depth,Y,X)\n",
109		" groups: \n",
110		"\n"
111		]
112		}
113		],
114		"source": [
115		"f = netCDF4.Dataset('data/rtofs_glo_3dz_f006_6hrly_reg3.nc')\n",
116		"print(f) "
117		]
118		},
119		{
120		"cell_type": "markdown",
121		"metadata": {
122		"internals": {
123		"frag_helper": "fragment_end",
124		"frag_number": 4,
125		"slide_type": "subslide"
126		},
127		"slideshow": {
128		"slide_type": "slide"
129		}
130		},
131		"source": [
132		"## Access a netCDF variable\n",
133		"- variable objects stored by name in `variables` dict.\n",
134		"- print the variable yields summary info (including all the attributes).\n",
135		"- no actual data read yet (just have a reference to the variable object with metadata)."
136		]
137		},
138		{
139		"cell_type": "code",
140		"execution_count": 7,
141		"metadata": {
142		"collapsed": false,
143		"internals": {
144		"frag_helper": "fragment_end",
145		"frag_number": 6,
146		"slide_helper": "subslide_end"
147		},
148		"slide_helper": "slide_end",
149		"slideshow": {
150		"slide_type": "fragment"
151		}
152		},
153		"outputs": [
154		{
155		"name": "stdout",
156		"output_type": "stream",
157		"text": [
158		"[u'MT', u'Date', u'Depth', u'Y', u'X', u'Latitude', u'Longitude', u'u', u'v', u'temperature', u'salinity']\n",
159		"<type 'netCDF4._netCDF4.Variable'>\n",
160		"float32 temperature(MT, Depth, Y, X)\n",
161		" coordinates: Longitude Latitude Date\n",
162		" standard_name: sea_water_potential_temperature\n",
163		" units: degC\n",
164		" _FillValue: 1.26765e+30\n",
165		" valid_range: [ -5.07860279 11.14989948]\n",
166		" long_name: temp [90.9H]\n",
167		"unlimited dimensions: MT\n",
168		"current shape = (1, 10, 850, 712)\n",
169		"filling on\n"
170		]
171		}
172		],
173		"source": [
174		"print(f.variables.keys()) # get all variable names\n",
175		"temp = f.variables['temperature'] # temperature variable\n",
176		"print(temp) "
177		]
178		},
179		{
180		"cell_type": "markdown",
181		"metadata": {
182		"internals": {
183		"frag_helper": "fragment_end",
184		"frag_number": 6,
185		"slide_type": "subslide"
186		},
187		"slideshow": {
188		"slide_type": "slide"
189		}
190		},
191		"source": [
192		"## List the Dimensions\n",
193		"\n",
194		"- All variables in a netCDF file have an associated shape, specified by a list of dimensions.\n",
195		"- Let's list all the dimensions in this netCDF file.\n",
196		"- Note that the `MT` dimension is special (`unlimited`), which means it can be appended to."
197		]
198		},
199		{
200		"cell_type": "code",
201		"execution_count": 8,
202		"metadata": {
203		"collapsed": false,
204		"internals": {
205		"frag_helper": "fragment_end",
206		"frag_number": 8
207		},
208		"slideshow": {
209		"slide_type": "fragment"
210		}
211		},
212		"outputs": [
213		{
214		"name": "stdout",
215		"output_type": "stream",
216		"text": [
217		"(u'MT', <type 'netCDF4._netCDF4.Dimension'> (unlimited): name = 'MT', size = 1\n",
218		")\n",
219		"(u'Y', <type 'netCDF4._netCDF4.Dimension'>: name = 'Y', size = 850\n",
220		")\n",
221		"(u'X', <type 'netCDF4._netCDF4.Dimension'>: name = 'X', size = 712\n",
222		")\n",
223		"(u'Depth', <type 'netCDF4._netCDF4.Dimension'>: name = 'Depth', size = 10\n",
224		")\n"
225		]
226		}
227		],
228		"source": [
229		"for d in f.dimensions.items():\n",
230		" print(d)"
231		]
232		},
233		{
234		"cell_type": "markdown",
235		"metadata": {
236		"internals": {
237		"frag_helper": "fragment_end",
238		"frag_number": 9
239		},
240		"slideshow": {
241		"slide_type": "fragment"
242		}
243		},
244		"source": [
245		"Each variable has a `dimensions` and a `shape` attribute."
246		]
247		},
248		{
249		"cell_type": "code",
250		"execution_count": 9,
251		"metadata": {
252		"collapsed": false,
253		"internals": {
254		"frag_helper": "fragment_end",
255		"frag_number": 10
256		},
257		"slideshow": {
258		"slide_type": "fragment"
259		}
260		},
261		"outputs": [
262		{
263		"data": {
264		"text/plain": [
265		"(u'MT', u'Depth', u'Y', u'X')"
266		]
267		},
268		"execution_count": 9,
269		"metadata": {},
270		"output_type": "execute_result"
271		}
272		],
273		"source": [
274		"temp.dimensions"
275		]
276		},
277		{
278		"cell_type": "code",
279		"execution_count": 10,
280		"metadata": {
281		"collapsed": false,
282		"internals": {
283		"frag_helper": "fragment_end",
284		"frag_number": 11,
285		"slide_helper": "subslide_end"
286		},
287		"slide_helper": "slide_end",
288		"slideshow": {
289		"slide_type": "fragment"
290		}
291		},
292		"outputs": [
293		{
294		"data": {
295		"text/plain": [
296		"(1, 10, 850, 712)"
297		]
298		},
299		"execution_count": 10,
300		"metadata": {},
301		"output_type": "execute_result"
302		}
303		],
304		"source": [
305		"temp.shape"
306		]
307		},
308		{
309		"cell_type": "markdown",
310		"metadata": {
311		"internals": {
312		"frag_helper": "fragment_end",
313		"frag_number": 11,
314		"slide_type": "subslide"
315		},
316		"slideshow": {
317		"slide_type": "slide"
318		}
319		},
320		"source": [
321		"### Each dimension typically has a variable associated with it (called a coordinate variable).\n",
322		"- Coordinate variables are 1D variables that have the same name as dimensions.\n",
323		"- Coordinate variables and auxiliary coordinate variables (named by the coordinates attribute) locate values in time and space."
324		]
325		},
326		{
327		"cell_type": "code",
328		"execution_count": 11,
329		"metadata": {
330		"collapsed": false,
331		"internals": {
332		"frag_helper": "fragment_end",
333		"frag_number": 13,
334		"slide_helper": "subslide_end"
335		},
336		"slide_helper": "slide_end",
337		"slideshow": {
338		"slide_type": "fragment"
339		}
340		},
341		"outputs": [
342		{
343		"name": "stdout",
344		"output_type": "stream",
345		"text": [
346		"<type 'netCDF4._netCDF4.Variable'>\n",
347		"float64 MT(MT)\n",
348		" long_name: time\n",
349		" units: days since 1900-12-31 00:00:00\n",
350		" calendar: standard\n",
351		" axis: T\n",
352		"unlimited dimensions: MT\n",
353		"current shape = (1,)\n",
354		"filling on, default _FillValue of 9.96920996839e+36 used\n",
355		"\n",
356		"<type 'netCDF4._netCDF4.Variable'>\n",
357		"int32 X(X)\n",
358		" point_spacing: even\n",
359		" axis: X\n",
360		"unlimited dimensions: \n",
361		"current shape = (712,)\n",
362		"filling on, default _FillValue of -2147483647 used\n",
363		"\n"
364		]
365		}
366		],
367		"source": [
368		"mt = f.variables['MT']\n",
369		"depth = f.variables['Depth']\n",
370		"x,y = f.variables['X'], f.variables['Y']\n",
371		"print(mt)\n",
372		"print(x) "
373		]
374		},
375		{
376		"cell_type": "markdown",
377		"metadata": {
378		"internals": {
379		"frag_helper": "fragment_end",
380		"frag_number": 13,
381		"slide_type": "subslide"
382		},
383		"slideshow": {
384		"slide_type": "slide"
385		}
386		},
387		"source": [
388		"## Accessing data from a netCDF variable object\n",
389		"\n",
390		"- netCDF variables objects behave much like numpy arrays.\n",
391		"- slicing a netCDF variable object returns a numpy array with the data.\n",
392		"- Boolean array and integer sequence indexing behaves differently for netCDF variables than for numpy arrays. Only 1-d boolean arrays and integer sequences are allowed, and these indices work independently along each dimension (similar to the way vector subscripts work in fortran)."
393		]
394		},
395		{
396		"cell_type": "code",
397		"execution_count": 12,
398		"metadata": {
399		"collapsed": false,
400		"internals": {
401		"frag_helper": "fragment_end",
402		"frag_number": 15
403		},
404		"slideshow": {
405		"slide_type": "fragment"
406		}
407		},
408		"outputs": [
409		{
410		"name": "stdout",
411		"output_type": "stream",
412		"text": [
413		"[ 41023.25]\n"
414		]
415		}
416		],
417		"source": [
418		"time = mt[:] # Reads the netCDF variable MT, array of one element\n",
419		"print(time) "
420		]
421		},
422		{
423		"cell_type": "code",
424		"execution_count": 13,
425		"metadata": {
426		"collapsed": false,
427		"internals": {
428		"frag_helper": "fragment_end",
429		"frag_number": 16
430		},
431		"slideshow": {
432		"slide_type": "fragment"
433		}
434		},
435		"outputs": [
436		{
437		"name": "stdout",
438		"output_type": "stream",
439		"text": [
440		"[ 0. 100. 200. 400. 700. 1000. 2000. 3000. 4000. 5000.]\n"
441		]
442		}
443		],
444		"source": [
445		"dpth = depth[:] # examine depth array\n",
446		"print(dpth) "
447		]
448		},
449		{
450		"cell_type": "code",
451		"execution_count": 14,
452		"metadata": {
453		"collapsed": false,
454		"internals": {
455		"frag_helper": "fragment_end",
456		"frag_number": 17,
457		"slide_helper": "subslide_end"
458		},
459		"slide_helper": "slide_end",
460		"slideshow": {
461		"slide_type": "fragment"
462		}
463		},
464		"outputs": [
465		{
466		"name": "stdout",
467		"output_type": "stream",
468		"text": [
469		"shape of temp variable: (1, 10, 850, 712)\n",
470		"shape of temp slice: (6, 425, 356)\n"
471		]
472		}
473		],
474		"source": [
475		"xx,yy = x[:],y[:]\n",
476		"print('shape of temp variable: %s' % repr(temp.shape))\n",
477		"tempslice = temp[0, dpth > 400, yy > yy.max()/2, xx > xx.max()/2]\n",
478		"print('shape of temp slice: %s' % repr(tempslice.shape))"
479		]
480		},
481		{
482		"cell_type": "markdown",
483		"metadata": {
484		"internals": {
485		"frag_helper": "fragment_end",
486		"frag_number": 17,
487		"slide_type": "subslide"
488		},
489		"slideshow": {
490		"slide_type": "slide"
491		}
492		},
493		"source": [
494		"## What is the sea surface temperature and salinity at 50N, 140W?\n",
495		"### Finding the latitude and longitude indices of 50N, 140W\n",
496		"\n",
497		"- The `X` and `Y` dimensions don't look like longitudes and latitudes\n",
498		"- Use the auxilary coordinate variables named in the `coordinates` variable attribute, `Latitude` and `Longitude`"
499		]
500		},
501		{
502		"cell_type": "code",
503		"execution_count": 15,
504		"metadata": {
505		"collapsed": false,
506		"internals": {
507		"frag_helper": "fragment_end",
508		"frag_number": 19
509		},
510		"slideshow": {
511		"slide_type": "fragment"
512		}
513		},
514		"outputs": [
515		{
516		"name": "stdout",
517		"output_type": "stream",
518		"text": [
519		"<type 'netCDF4._netCDF4.Variable'>\n",
520		"float32 Latitude(Y, X)\n",
521		" standard_name: latitude\n",
522		" units: degrees_north\n",
523		"unlimited dimensions: \n",
524		"current shape = (850, 712)\n",
525		"filling on, default _FillValue of 9.96920996839e+36 used\n",
526		"\n"
527		]
528		}
529		],
530		"source": [
531		"lat, lon = f.variables['Latitude'], f.variables['Longitude']\n",
532		"print(lat)"
533		]
534		},
535		{
536		"cell_type": "markdown",
537		"metadata": {
538		"internals": {
539		"frag_helper": "fragment_end",
540		"frag_number": 20,
541		"slide_helper": "subslide_end"
542		},
543		"slide_helper": "slide_end",
544		"slideshow": {
545		"slide_type": "fragment"
546		}
547		},
548		"source": [
549		"Aha! So we need to find array indices `iy` and `ix` such that `Latitude[iy, ix]` is close to 50.0 and `Longitude[iy, ix]` is close to -140.0 ..."
550		]
551		},
552		{
553		"cell_type": "code",
554		"execution_count": 16,
555		"metadata": {
556		"collapsed": false,
557		"internals": {
558		"frag_helper": "fragment_end",
559		"frag_number": 20,
560		"slide_type": "subslide"
561		},
562		"slideshow": {
563		"slide_type": "slide"
564		}
565		},
566		"outputs": [],
567		"source": [
568		"# extract lat/lon values (in degrees) to numpy arrays\n",
569		"latvals = lat[:]; lonvals = lon[:] \n",
570		"# a function to find the index of the point closest pt\n",
571		"# (in squared distance) to give lat/lon value.\n",
572		"def getclosest_ij(lats,lons,latpt,lonpt):\n",
573		" # find squared distance of every point on grid\n",
574		" dist_sq = (lats-latpt)2 + (lons-lonpt)2 \n",
575		" # 1D index of minimum dist_sq element\n",
576		" minindex_flattened = dist_sq.argmin() \n",
577		" # Get 2D index for latvals and lonvals arrays from 1D index\n",
578		" return np.unravel_index(minindex_flattened, lats.shape)\n",
579		"iy_min, ix_min = getclosest_ij(latvals, lonvals, 50., -140)"
580		]
581		},
582		{
583		"cell_type": "markdown",
584		"metadata": {
585		"internals": {
586		"frag_helper": "fragment_end",
587		"frag_number": 22
588		},
589		"slideshow": {
590		"slide_type": "fragment"
591		}
592		},
593		"source": [
594		"### Now we have all the information we need to find our answer.\n"
595		]
596		},
597		{
598		"cell_type": "markdown",
599		"metadata": {
600		"internals": {
601		"frag_helper": "fragment_end",
602		"frag_number": 23
603		},
604		"slideshow": {
605		"slide_type": "fragment"
606		}
607		},
608		"source": [
609		"```\n",
610		"\|----------+--------\|\n",
611		"\| Variable \| Index \|\n",
612		"\|----------+--------\|\n",
613		"\| MT \| 0 \|\n",
614		"\| Depth \| 0 \|\n",
615		"\| Y \| iy_min \|\n",
616		"\| X \| ix_min \|\n",
617		"\|----------+--------\|\n",
618		"```"
619		]
620		},
621		{
622		"cell_type": "markdown",
623		"metadata": {
624		"internals": {
625		"frag_helper": "fragment_end",
626		"frag_number": 24
627		},
628		"slideshow": {
629		"slide_type": "fragment"
630		}
631		},
632		"source": [
633		"### What is the sea surface temperature and salinity at the specified point?"
634		]
635		},
636		{
637		"cell_type": "code",
638		"execution_count": 17,
639		"metadata": {
640		"collapsed": false,
641		"internals": {
642		"frag_helper": "fragment_end",
643		"frag_number": 25,
644		"slide_helper": "subslide_end"
645		},
646		"slide_helper": "slide_end",
647		"slideshow": {
648		"slide_type": "fragment"
649		}
650		},
651		"outputs": [
652		{
653		"name": "stdout",
654		"output_type": "stream",
655		"text": [
656		" 6.4631 degC\n",
657		"32.6572 psu\n"
658		]
659		}
660		],
661		"source": [
662		"sal = f.variables['salinity']\n",
663		"# Read values out of the netCDF file for temperature and salinity\n",
664		"print('%7.4f %s' % (temp[0,0,iy_min,ix_min], temp.units))\n",
665		"print('%7.4f %s' % (sal[0,0,iy_min,ix_min], sal.units))"
666		]
667		},
668		{
669		"cell_type": "markdown",
670		"metadata": {
671		"internals": {
672		"frag_helper": "fragment_end",
673		"frag_number": 25,
674		"slide_type": "subslide"
675		},
676		"slideshow": {
677		"slide_type": "slide"
678		}
679		},
680		"source": [
681		"## Remote data access via openDAP\n",
682		"\n",
683		"- Remote data can be accessed seamlessly with the netcdf4-python API\n",
684		"- Access happens via the DAP protocol and DAP servers, such as TDS.\n",
685		"- many formats supported, like GRIB, are supported \"under the hood\"."
686		]
687		},
688		{
689		"cell_type": "markdown",
690		"metadata": {
691		"internals": {
692		"frag_helper": "fragment_end",
693		"frag_number": 27
694		},
695		"slideshow": {
696		"slide_type": "fragment"
697		}
698		},
699		"source": [
700		"The following example showcases some nice netCDF features:\n",
701		"\n",
702		"1. We are seamlessly accessing remote data, from a TDS server.\n",
703		"2. We are seamlessly accessing GRIB2 data, as if it were netCDF data.\n",
704		"3. We are generating metadata on-the-fly."
705		]
706		},
707		{
708		"cell_type": "code",
709		"execution_count": 19,
710		"metadata": {
711		"collapsed": false,
712		"internals": {
713		"frag_helper": "fragment_end",
714		"frag_number": 28,
715		"slide_helper": "subslide_end"
716		},
717		"slide_helper": "slide_end",
718		"slideshow": {
719		"slide_type": "fragment"
720		}
721		},
722		"outputs": [
723		{
724		"name": "stdout",
725		"output_type": "stream",
726		"text": [
727		"http://thredds.ucar.edu/thredds/dodsC/grib/NCEP/GFS/Global_0p5deg/GFS_Global_0p5deg_20150711_0600.grib2/GC\n"
728		]
729		}
730		],
731		"source": [
732		"import datetime\n",
733		"date = datetime.datetime.now()\n",
734		"# build URL for latest synoptic analysis time\n",
735		"URL = 'http://thredds.ucar.edu/thredds/dodsC/grib/NCEP/GFS/Global_0p5deg/GFS_Global_0p5deg_%04i%02i%02i_%02i%02i.grib2/GC' %\\\n",
736		"(date.year,date.month,date.day,6*(date.hour//6),0)\n",
737		"# keep moving back 6 hours until a valid URL found\n",
738		"validURL = False; ncount = 0\n",
739		"while (not validURL and ncount < 10):\n",
740		" print(URL)\n",
741		" try:\n",
742		" gfs = netCDF4.Dataset(URL)\n",
743		" validURL = True\n",
744		" except RuntimeError:\n",
745		" date -= datetime.timedelta(hours=6)\n",
746		" ncount += 1 "
747		]
748		},
749		{
750		"cell_type": "code",
751		"execution_count": 20,
752		"metadata": {
753		"collapsed": false,
754		"internals": {
755		"frag_helper": "fragment_end",
756		"frag_number": 28,
757		"slide_helper": "subslide_end",
758		"slide_type": "subslide"
759		},
760		"slide_helper": "slide_end",
761		"slideshow": {
762		"slide_type": "slide"
763		}
764		},
765		"outputs": [
766		{
767		"name": "stdout",
768		"output_type": "stream",
769		"text": [
770		"<type 'netCDF4._netCDF4.Variable'>\n",
771		"float32 Temperature_surface(time2, lat, lon)\n",
772		" long_name: Temperature @ Ground or water surface\n",
773		" units: K\n",
774		" abbreviation: TMP\n",
775		" missing_value: nan\n",
776		" grid_mapping: LatLon_Projection\n",
777		" coordinates: reftime time2 lat lon \n",
778		" Grib_Variable_Id: VAR_0-0-0_L1\n",
779		" Grib2_Parameter: [0 0 0]\n",
780		" Grib2_Parameter_Discipline: Meteorological products\n",
781		" Grib2_Parameter_Category: Temperature\n",
782		" Grib2_Parameter_Name: Temperature\n",
783		" Grib2_Level_Type: Ground or water surface\n",
784		" Grib2_Generating_Process_Type: Forecast\n",
785		"unlimited dimensions: \n",
786		"current shape = (93, 361, 720)\n",
787		"filling off\n",
788		"\n",
789		"<type 'netCDF4._netCDF4.Variable'>\n",
790		"float64 time2(time2)\n",
791		" units: Hour since 2015-07-11T06:00:00Z\n",
792		" standard_name: time\n",
793		" long_name: GRIB forecast or observation time\n",
794		" calendar: proleptic_gregorian\n",
795		" _CoordinateAxisType: Time\n",
796		"unlimited dimensions: \n",
797		"current shape = (93,)\n",
798		"filling off\n",
799		"\n",
800		"<type 'netCDF4._netCDF4.Variable'>\n",
801		"float32 lat(lat)\n",
802		" units: degrees_north\n",
803		" _CoordinateAxisType: Lat\n",
804		"unlimited dimensions: \n",
805		"current shape = (361,)\n",
806		"filling off\n",
807		"\n",
808		"<type 'netCDF4._netCDF4.Variable'>\n",
809		"float32 lon(lon)\n",
810		" units: degrees_east\n",
811		" _CoordinateAxisType: Lon\n",
812		"unlimited dimensions: \n",
813		"current shape = (720,)\n",
814		"filling off\n",
815		"\n"
816		]
817		}
818		],
819		"source": [
820		"# Look at metadata for a specific variable\n",
821		"# gfs.variables.keys() will show all available variables.\n",
822		"sfctmp = gfs.variables['Temperature_surface']\n",
823		"# get info about sfctmp\n",
824		"print(sfctmp)\n",
825		"# print coord vars associated with this variable\n",
826		"for dname in sfctmp.dimensions: \n",
827		" print(gfs.variables[dname])"
828		]
829		},
830		{
831		"cell_type": "markdown",
832		"metadata": {
833		"internals": {
834		"frag_helper": "fragment_end",
835		"frag_number": 28,
836		"slide_type": "subslide"
837		},
838		"slideshow": {
839		"slide_type": "slide"
840		}
841		},
842		"source": [
843		"##Missing values\n",
844		"- when `data == var.missing_value` somewhere, a masked array is returned.\n",
845		"- illustrate with soil moisture data (only defined over land)\n",
846		"- white areas on plot are masked values over water."
847		]
848		},
849		{
850		"cell_type": "code",
851		"execution_count": 21,
852		"metadata": {
853		"collapsed": false,
854		"internals": {
855		"frag_helper": "fragment_end",
856		"frag_number": 31
857		},
858		"slideshow": {
859		"slide_type": "fragment"
860		}
861		},
862		"outputs": [
863		{
864		"name": "stdout",
865		"output_type": "stream",
866		"text": [
867		"shape=(361, 720), type=<class 'numpy.ma.core.MaskedArray'>, missing_value=nan\n"
868		]
869		},
870		{
871		"data": {
872		"image/png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAAD7CAYAAAB37B+tAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIGSOCLkgCELmiJALgiBkjgi5IAhC5vx/oWJ9OHx0YTwAAAAASUVORK5CYII=\n",
873		"text/plain": [
874		"<matplotlib.figure.Figure at 0x1125c5a90>"
875		]
876		},
877		"metadata": {},
878		"output_type": "display_data"
879		}
880		],
881		"source": [
882		"soilmvar = gfs.variables['Volumetric_Soil_Moisture_Content_depth_below_surface_layer']\n",
883		"# flip the data in latitude so North Hemisphere is up on the plot\n",
884		"soilm = soilmvar[0,0,::-1,:] \n",
885		"print('shape=%s, type=%s, missing_value=%s' % \\\n",
886		" (soilm.shape, type(soilm), soilmvar.missing_value))\n",
887		"import matplotlib.pyplot as plt\n",
888		"%matplotlib inline\n",
889		"cs = plt.contourf(soilm)"
890		]
891		},
892		{
893		"cell_type": "markdown",
894		"metadata": {
895		"internals": {
896		"frag_helper": "fragment_end",
897		"frag_number": 32,
898		"slide_helper": "subslide_end"
899		},
900		"slide_helper": "slide_end",
901		"slideshow": {
902		"slide_type": "fragment"
903		}
904		},
905		"source": [
906		"##Packed integer data\n",
907		"There is a similar feature for variables with `scale_factor` and `add_offset` attributes.\n",
908		"\n",
909		"- short integer data will automatically be returned as float data, with the scale and offset applied. "
910		]
911		},
912		{
913		"cell_type": "markdown",
914		"metadata": {
915		"internals": {
916		"frag_helper": "fragment_end",
917		"frag_number": 32,
918		"slide_type": "subslide"
919		},
920		"slideshow": {
921		"slide_type": "slide"
922		}
923		},
924		"source": [
925		"## Dealing with dates and times\n",
926		"- time variables usually measure relative to a fixed date using a certain calendar, with units specified like *`hours since YY:MM:DD hh-mm-ss`*.\n",
927		"- `num2date` and `date2num` convenience functions provided to convert between these numeric time coordinates and handy python datetime instances. \n",
928		"- `date2index` finds the time index corresponding to a datetime instance."
929		]
930		},
931		{
932		"cell_type": "code",
933		"execution_count": 22,
934		"metadata": {
935		"collapsed": false,
936		"internals": {
937		"frag_helper": "fragment_end",
938		"frag_number": 34
939		},
940		"slideshow": {
941		"slide_type": "fragment"
942		}
943		},
944		"outputs": [
945		{
946		"name": "stdout",
947		"output_type": "stream",
948		"text": [
949		"name of time dimension = time2\n",
950		"units = Hour since 2015-07-11T06:00:00Z, values = [ 0. 3. 6. 9. 12. 15. 18. 21. 24. 27. 30. 33.\n",
951		" 36. 39. 42. 45. 48. 51. 54. 57. 60. 63. 66. 69.\n",
952		" 72. 75. 78. 81. 84. 87. 90. 93. 96. 99. 102. 105.\n",
953		" 108. 111. 114. 117. 120. 123. 126. 129. 132. 135. 138. 141.\n",
954		" 144. 147. 150. 153. 156. 159. 162. 165. 168. 171. 174. 177.\n",
955		" 180. 183. 186. 189. 192. 195. 198. 201. 204. 207. 210. 213.\n",
956		" 216. 219. 222. 225. 228. 231. 234. 237. 240. 252. 264. 276.\n",
957		" 288. 300. 312. 324. 336. 348. 360. 372. 384.]\n"
958		]
959		}
960		],
961		"source": [
962		"from netCDF4 import num2date, date2num, date2index\n",
963		"timedim = sfctmp.dimensions[0] # time dim name\n",
964		"print('name of time dimension = %s' % timedim)\n",
965		"times = gfs.variables[timedim] # time coord var\n",
966		"print('units = %s, values = %s' % (times.units, times[:]))"
967		]
968		},
969		{
970		"cell_type": "code",
971		"execution_count": 23,
972		"metadata": {
973		"collapsed": false,
974		"internals": {
975		"frag_helper": "fragment_end",
976		"frag_number": 35,
977		"slide_helper": "subslide_end"
978		},
979		"slide_helper": "slide_end",
980		"slideshow": {
981		"slide_type": "fragment"
982		}
983		},
984		"outputs": [
985		{
986		"name": "stdout",
987		"output_type": "stream",
988		"text": [
989		"['2015-07-11 06:00:00', '2015-07-11 09:00:00', '2015-07-11 12:00:00', '2015-07-11 15:00:00', '2015-07-11 18:00:00', '2015-07-11 21:00:00', '2015-07-12 00:00:00', '2015-07-12 03:00:00', '2015-07-12 06:00:00', '2015-07-12 09:00:00']\n"
990		]
991		}
992		],
993		"source": [
994		"dates = num2date(times[:], times.units)\n",
995		"print([date.strftime('%Y-%m-%d %H:%M:%S') for date in dates[:10]]) # print only first ten..."
996		]
997		},
998		{
999		"cell_type": "markdown",
1000		"metadata": {
1001		"internals": {
1002		"frag_helper": "fragment_end",
1003		"frag_number": 35,
1004		"slide_type": "subslide"
1005		},
1006		"slideshow": {
1007		"slide_type": "slide"
1008		}
1009		},
1010		"source": [
1011		"###Get index associated with a specified date, extract forecast data for that date."
1012		]
1013		},
1014		{
1015		"cell_type": "code",
1016		"execution_count": 24,
1017		"metadata": {
1018		"collapsed": false,
1019		"internals": {
1020		"frag_helper": "fragment_end",
1021		"frag_number": 37
1022		},
1023		"slideshow": {
1024		"slide_type": "fragment"
1025		}
1026		},
1027		"outputs": [
1028		{
1029		"name": "stdout",
1030		"output_type": "stream",
1031		"text": [
1032		"2015-07-14 07:22:39.579246\n",
1033		"index = 24, date = 2015-07-14 06:00:00\n"
1034		]
1035		}
1036		],
1037		"source": [
1038		"from datetime import datetime, timedelta\n",
1039		"date = datetime.now() + timedelta(days=3)\n",
1040		"print(date)\n",
1041		"ntime = date2index(date,times,select='nearest')\n",
1042		"print('index = %s, date = %s' % (ntime, dates[ntime]))"
1043		]
1044		},
1045		{
1046		"cell_type": "markdown",
1047		"metadata": {
1048		"internals": {
1049		"frag_helper": "fragment_end",
1050		"frag_number": 38
1051		},
1052		"slideshow": {
1053		"slide_type": "fragment"
1054		}
1055		},
1056		"source": [
1057		"###Get temp forecast for Boulder (near 40N, -105W)\n",
1058		"- use function `getcloses_ij` we created before..."
1059		]
1060		},
1061		{
1062		"cell_type": "code",
1063		"execution_count": 25,
1064		"metadata": {
1065		"collapsed": false,
1066		"internals": {
1067		"frag_helper": "fragment_end",
1068		"frag_number": 39,
1069		"slide_helper": "subslide_end"
1070		},
1071		"slide_helper": "slide_end",
1072		"slideshow": {
1073		"slide_type": "fragment"
1074		}
1075		},
1076		"outputs": [
1077		{
1078		"name": "stdout",
1079		"output_type": "stream",
1080		"text": [
1081		"Boulder forecast valid at 2015-07-14 06:00:00 UTC = 296.8 K\n"
1082		]
1083		}
1084		],
1085		"source": [
1086		"lats, lons = gfs.variables['lat'][:], gfs.variables['lon'][:]\n",
1087		"# lats, lons are 1-d. Make them 2-d using numpy.meshgrid.\n",
1088		"lons, lats = np.meshgrid(lons,lats)\n",
1089		"j, i = getclosest_ij(lats,lons,40,-105)\n",
1090		"fcst_temp = sfctmp[ntime,j,i]\n",
1091		"print('Boulder forecast valid at %s UTC = %5.1f %s' % \\\n",
1092		" (dates[ntime],fcst_temp,sfctmp.units))"
1093		]
1094		},
1095		{
1096		"cell_type": "markdown",
1097		"metadata": {
1098		"internals": {
1099		"frag_helper": "fragment_end",
1100		"frag_number": 39,
1101		"slide_type": "subslide"
1102		},
1103		"slideshow": {
1104		"slide_type": "slide"
1105		}
1106		},
1107		"source": [
1108		"##Simple multi-file aggregation\n",
1109		"\n",
1110		"What if you have a bunch of netcdf files, each with data for a different year, and you want to access all the data as if it were in one file?"
1111		]
1112		},
1113		{
1114		"cell_type": "code",
1115		"execution_count": 26,
1116		"metadata": {
1117		"collapsed": false,
1118		"internals": {
1119		"frag_helper": "fragment_end",
1120		"frag_number": 41
1121		},
1122		"slideshow": {
1123		"slide_type": "fragment"
1124		}
1125		},
1126		"outputs": [
1127		{
1128		"name": "stdout",
1129		"output_type": "stream",
1130		"text": [
1131		"-rw-r--r-- 1 jwhitaker staff 8985332 Jul 10 06:43 data/prmsl.2000.nc\r\n",
1132		"-rw-r--r-- 1 jwhitaker staff 8968789 Jul 10 06:43 data/prmsl.2001.nc\r\n",
1133		"-rw-r--r-- 1 jwhitaker staff 8972796 Jul 10 06:43 data/prmsl.2002.nc\r\n",
1134		"-rw-r--r-- 1 jwhitaker staff 8974435 Jul 10 06:43 data/prmsl.2003.nc\r\n",
1135		"-rw-r--r-- 1 jwhitaker staff 8997438 Jul 10 06:43 data/prmsl.2004.nc\r\n",
1136		"-rw-r--r-- 1 jwhitaker staff 8976678 Jul 10 06:43 data/prmsl.2005.nc\r\n",
1137		"-rw-r--r-- 1 jwhitaker staff 8969714 Jul 10 06:43 data/prmsl.2006.nc\r\n",
1138		"-rw-r--r-- 1 jwhitaker staff 8974360 Jul 10 06:43 data/prmsl.2007.nc\r\n",
1139		"-rw-r--r-- 1 jwhitaker staff 8994260 Jul 10 06:43 data/prmsl.2008.nc\r\n",
1140		"-rw-r--r-- 1 jwhitaker staff 8974678 Jul 10 06:43 data/prmsl.2009.nc\r\n",
1141		"-rw-r--r-- 1 jwhitaker staff 8970732 Jul 10 06:43 data/prmsl.2010.nc\r\n",
1142		"-rw-r--r-- 1 jwhitaker staff 8976285 Jul 10 06:43 data/prmsl.2011.nc\r\n"
1143		]
1144		}
1145		],
1146		"source": [
1147		"!ls -l data/prmsl*nc"
1148		]
1149		},
1150		{
1151		"cell_type": "markdown",
1152		"metadata": {
1153		"internals": {
1154		"frag_helper": "fragment_end",
1155		"frag_number": 42
1156		},
1157		"slideshow": {
1158		"slide_type": "fragment"
1159		}
1160		},
1161		"source": [
1162		"`MFDataset` uses file globbing to patch together all the files into one big Dataset.\n",
1163		"You can also pass it a list of specific files.\n",
1164		"\n",
1165		"Limitations:\n",
1166		"\n",
1167		"- It can only aggregate the data along the leftmost dimension of each variable.\n",
1168		"- only works with `NETCDF3`, or `NETCDF4_CLASSIC` formatted files.\n",
1169		"- kind of slow."
1170		]
1171		},
1172		{
1173		"cell_type": "code",
1174		"execution_count": 27,
1175		"metadata": {
1176		"collapsed": false,
1177		"internals": {
1178		"frag_helper": "fragment_end",
1179		"frag_number": 43,
1180		"slide_helper": "subslide_end"
1181		},
1182		"slide_helper": "slide_end",
1183		"slideshow": {
1184		"slide_type": "fragment"
1185		}
1186		},
1187		"outputs": [
1188		{
1189		"name": "stdout",
1190		"output_type": "stream",
1191		"text": [
1192		"starting date = 2000-01-01 00:00:00\n",
1193		"ending date = 2011-12-31 00:00:00\n",
1194		"times shape = 4383\n",
1195		"prmsl dimensions = (u'time', u'lat', u'lon'), prmsl shape = (4383, 91, 180)\n"
1196		]
1197		}
1198		],
1199		"source": [
1200		"mf = netCDF4.MFDataset('data/prmsl*nc')\n",
1201		"times = mf.variables['time']\n",
1202		"dates = num2date(times[:],times.units)\n",
1203		"print('starting date = %s' % dates[0])\n",
1204		"print('ending date = %s'% dates[-1])\n",
1205		"prmsl = mf.variables['prmsl']\n",
1206		"print('times shape = %s' % times.shape)\n",
1207		"print('prmsl dimensions = %s, prmsl shape = %s' %\\\n",
1208		" (prmsl.dimensions, prmsl.shape))"
1209		]
1210		},
1211		{
1212		"cell_type": "markdown",
1213		"metadata": {
1214		"internals": {
1215		"frag_helper": "fragment_end",
1216		"frag_number": 43,
1217		"slide_type": "subslide"
1218		},
1219		"slideshow": {
1220		"slide_type": "slide"
1221		}
1222		},
1223		"source": [
1224		"## Closing your netCDF file\n",
1225		"\n",
1226		"It's good to close netCDF files, but not actually necessary when Dataset is open for read access only.\n"
1227		]
1228		},
1229		{
1230		"cell_type": "code",
1231		"execution_count": 28,
1232		"metadata": {
1233		"collapsed": false,
1234		"internals": {
1235		"frag_helper": "fragment_end",
1236		"frag_number": 45
1237		},
1238		"slideshow": {
1239		"slide_type": "fragment"
1240		}
1241		},
1242		"outputs": [],
1243		"source": [
1244		"f.close()\n",
1245		"gfs.close()"
1246		]
1247		},
1248		{
1249		"cell_type": "markdown",
1250		"metadata": {
1251		"internals": {
1252		"frag_helper": "fragment_end",
1253		"frag_number": 45,
1254		"slide_helper": "subslide_end"
1255		},
1256		"slide_helper": "slide_end",
1257		"slideshow": {
1258		"slide_type": "-"
1259		}
1260		},
1261		"source": [
1262		"##That's it!\n",
1263		"\n",
1264		"Now you're ready to start exploring your data interactively.\n",
1265		"\n",
1266		"To be continued with Writing netCDF data ...."
1267		]
1268		}
1269		],
1270		"metadata": {
1271		"celltoolbar": "Raw Cell Format",
1272		"kernelspec": {
1273		"display_name": "Python 2",
1274		"language": "python",
1275		"name": "python2"
1276		},
1277		"language_info": {
1278		"codemirror_mode": {
1279		"name": "ipython",
1280		"version": 2
1281		},
1282		"file_extension": ".py",
1283		"mimetype": "text/x-python",
1284		"name": "python",
1285		"nbconvert_exporter": "python",
1286		"pygments_lexer": "ipython2",
1287		"version": "2.7.9"
1288		}
1289		},
1290		"nbformat": 4,
1291		"nbformat_minor": 0
1292		}

+37

-37

examples/tutorial.py less more

79	79
80	80	print(rootgrp.variables)
81	81
82		import numpy
	82	import numpy as np
83	83	# no unlimited dimension, just assign to slice.
84		lats = numpy.arange(-90,91,2.5)
85		lons = numpy.arange(-180,180,2.5)
	84	lats = np.arange(-90,91,2.5)
	85	lons = np.arange(-180,180,2.5)
86	86	latitudes[:] = lats
87	87	longitudes[:] = lons
88	88	print('latitudes =\n',latitudes[:])

122	122	f = Dataset('mftest'+repr(nfile)+'.nc','w',format='NETCDF4_CLASSIC')
123	123	f.createDimension('x',None)
124	124	x = f.createVariable('x','i',('x',))
125		x[0:10] = numpy.arange(nfile10,10(nfile+1))
	125	x[0:10] = np.arange(nfile10,10(nfile+1))
126	126	f.close()
127	127	# now read all those files in at once, in one Dataset.
128	128	from netCDF4 import MFDataset

133	133	f = Dataset('complex.nc','w')
134	134	size = 3 # length of 1-d complex array
135	135	# create sample complex data.
136		datac = numpy.exp(1j*(1.+numpy.linspace(0, numpy.pi, size)))
	136	datac = np.exp(1j*(1.+np.linspace(0, np.pi, size)))
137	137	print(datac.dtype)
138	138	# create complex128 compound data type.
139		complex128 = numpy.dtype([('real',numpy.float64),('imag',numpy.float64)])
	139	complex128 = np.dtype([('real',np.float64),('imag',np.float64)])
140	140	complex128_t = f.createCompoundType(complex128,'complex128')
141	141	# create a variable with this data type, write some data to it.
142	142	f.createDimension('x_dim',None)
143	143	v = f.createVariable('cmplx_var',complex128_t,'x_dim')
144		data = numpy.empty(size,complex128) # numpy structured array
	144	data = np.empty(size,complex128) # numpy structured array
145	145	data['real'] = datac.real; data['imag'] = datac.imag
146	146	v[:] = data
147	147	# close and reopen the file, check the contents.

155	155	print(v.shape)
156	156	datain = v[:] # read in all the data into a numpy structured array
157	157	# create an empty numpy complex array
158		datac2 = numpy.empty(datain.shape,numpy.complex128)
	158	datac2 = np.empty(datain.shape,np.complex128)
159	159	# .. fill it with contents of structured array.
160	160	datac2.real = datain['real']
161	161	datac2.imag = datain['imag']

167	167	# create an unlimited dimension call 'station'
168	168	f.createDimension('station',None)
169	169	# define a compound data type (can contain arrays, or nested compound types).
170		winddtype = numpy.dtype([('speed','f4'),('direction','i4')])
171		statdtype = numpy.dtype([('latitude', 'f4'), ('longitude', 'f4'),
172		('surface_wind',winddtype),
173		('temp_sounding','f4',10),('press_sounding','i4',10),
174		('location_name','S12')])
	170	winddtype = np.dtype([('speed','f4'),('direction','i4')])
	171	statdtype = np.dtype([('latitude', 'f4'), ('longitude', 'f4'),
	172	('surface_wind',winddtype),
	173	('temp_sounding','f4',10),('press_sounding','i4',10),
	174	('location_name','S12')])
175	175	# use this data type definitions to create a compound data types
176	176	# called using the createCompoundType Dataset method.
177	177	# create a compound type for vector wind which will be nested inside

180	180	# now that wind_data_t is defined, create the station data type.
181	181	station_data_t = f.createCompoundType(statdtype,'station_data')
182	182	# create nested compound data types to hold the units variable attribute.
183		winddtype_units = numpy.dtype([('speed','S12'),('direction','S12')])
184		statdtype_units = numpy.dtype([('latitude', 'S12'), ('longitude', 'S12'),
185		('surface_wind',winddtype_units),
186		('temp_sounding','S12'),
187		('location_name','S12'),
188		('press_sounding','S12')])
	183	winddtype_units = np.dtype([('speed','S12'),('direction','S12')])
	184	statdtype_units = np.dtype([('latitude', 'S12'), ('longitude', 'S12'),
	185	('surface_wind',winddtype_units),
	186	('temp_sounding','S12'),
	187	('location_name','S12'),
	188	('press_sounding','S12')])
189	189	# create the wind_data_units type first, since it will nested inside
190	190	# the station_data_units data type.
191	191	wind_data_units_t = f.createCompoundType(winddtype_units,'wind_data_units')

194	194	# create a variable of of type 'station_data_t'
195	195	statdat = f.createVariable('station_obs', station_data_t, ('station',))
196	196	# create a numpy structured array, assign data to it.
197		data = numpy.empty(1,statdtype)
	197	data = np.empty(1,statdtype)
198	198	data['latitude'] = 40.
199	199	data['longitude'] = -105.
200	200	data['surface_wind']['speed'] = 12.5

206	206	statdat[0] = data
207	207	# or just assign a tuple of values to variable slice
208	208	# (will automatically be converted to a structured array).
209		statdat[1] = numpy.array((40.78,-73.99,(-12.5,90),
	209	statdat[1] = np.array((40.78,-73.99,(-12.5,90),
210	210	(290.2,282.5,279.,277.9,276.,266.,264.1,260.,255.5,243.),
211	211	range(900,400,-50),'New York, NY'),data.dtype)
212	212	print(f.cmptypes)
213		windunits = numpy.empty(1,winddtype_units)
214		stationobs_units = numpy.empty(1,statdtype_units)
	213	windunits = np.empty(1,winddtype_units)
	214	stationobs_units = np.empty(1,statdtype_units)
215	215	windunits['speed'] = 'm/s'
216	216	windunits['direction'] = 'degrees'
217	217	stationobs_units['latitude'] = 'degrees N'

234	234	f.close()
235	235
236	236	f = Dataset('tst_vlen.nc','w')
237		vlen_t = f.createVLType(numpy.int32, 'phony_vlen')
	237	vlen_t = f.createVLType(np.int32, 'phony_vlen')
238	238	x = f.createDimension('x',3)
239	239	y = f.createDimension('y',4)
240	240	vlvar = f.createVariable('phony_vlen_var', vlen_t, ('y','x'))
241	241	import random
242		data = numpy.empty(len(y)*len(x),object)
	242	data = np.empty(len(y)*len(x),object)
243	243	for n in range(len(y)*len(x)):
244		data[n] = numpy.arange(random.randint(1,10),dtype='int32')+1
245		data = numpy.reshape(data,(len(y),len(x)))
	244	data[n] = np.arange(random.randint(1,10),dtype='int32')+1
	245	data = np.reshape(data,(len(y),len(x)))
246	246	vlvar[:] = data
247	247	print(vlvar)
248	248	print('vlen variable =\n',vlvar[:])

252	252	z = f.createDimension('z', 10)
253	253	strvar = f.createVariable('strvar',str,'z')
254	254	chars = '1234567890aabcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
255		data = numpy.empty(10,object)
	255	data = np.empty(10,object)
256	256	for n in range(10):
257	257	stringlen = random.randint(2,12)
258	258	data[n] = ''.join([random.choice(chars) for i in range(stringlen)])

269	269	u'Nimbostratus': 6, u'Cumulus': 4, u'Altostratus': 5, u'Cumulonimbus': 1,
270	270	u'Stratocumulus': 3}
271	271	# create the Enum type called 'cloud_t'.
272		cloud_type = f.createEnumType(numpy.uint8,'cloud_t',enum_dict)
	272	cloud_type = f.createEnumType(np.uint8,'cloud_t',enum_dict)
273	273	print(cloud_type)
274	274	time = f.createDimension('time',None)
275	275	# create a 1d variable of type 'cloud_type' called 'primary_clouds'.

294	294	nc.createDimension('nchars',3)
295	295	nc.createDimension('nstrings',None)
296	296	v = nc.createVariable('strings','S1',('nstrings','nchars'))
297		datain = numpy.array(['foo','bar'],dtype='S3')
	297	datain = np.array(['foo','bar'],dtype='S3')
298	298	v[:] = stringtochar(datain) # manual conversion to char array
299	299	print(v[:]) # data returned as char array
300	300	v._Encoding = 'ascii' # this enables automatic conversion

303	303	nc.close()
304	304	# strings in compound types
305	305	nc = Dataset('compoundstring_example.nc','w')
306		dtype = numpy.dtype([('observation', 'f4'),
307		('station_name','S12')])
	306	dtype = np.dtype([('observation', 'f4'),
	307	('station_name','S12')])
308	308	station_data_t = nc.createCompoundType(dtype,'station_data')
309	309	nc.createDimension('station',None)
310	310	statdat = nc.createVariable('station_obs', station_data_t, ('station',))
311		data = numpy.empty(2,station_data_t.dtype_view)
	311	data = np.empty(2,station_data_t.dtype_view)
312	312	data['observation'][:] = (123.,3.14)
313	313	data['station_name'][:] = ('Boulder','New York')
314	314	print(statdat.dtype) # strings actually stored as character arrays

324	324	# to disk when it is closed.
325	325	nc = Dataset('diskless_example.nc','w',diskless=True,persist=True)
326	326	d = nc.createDimension('x',None)
327		v = nc.createVariable('v',numpy.int32,'x')
328		v[0:5] = numpy.arange(5)
	327	v = nc.createVariable('v',np.int32,'x')
	328	v[0:5] = np.arange(5)
329	329	print(nc)
330	330	print(nc['v'][:])
331	331	nc.close() # file saved to disk

344	344	# used if format is NETCDF3 (ignored for NETCDF4/HDF5 files).
345	345	nc = Dataset('inmemory.nc', mode='w',memory=1028)
346	346	d = nc.createDimension('x',None)
347		v = nc.createVariable('v',numpy.int32,'x')
348		v[0:5] = numpy.arange(5)
	347	v = nc.createVariable('v',np.int32,'x')
	348	v[0:5] = np.arange(5)
349	349	nc_buf = nc.close() # close returns memoryview
350	350	print(type(nc_buf))
351	351	# save nc_buf to disk, read it back in and check.

-1206

~~examples/writing_netCDF.ipynb~~ less more

0		{
1		"cells": [
2		{
3		"cell_type": "markdown",
4		"metadata": {
5		"internals": {
6		"slide_type": "subslide"
7		},
8		"slideshow": {
9		"slide_type": "slide"
10		}
11		},
12		"source": [
13		"# Writing netCDF data\n",
14		"\n",
15		"Important Note: when running this notebook interactively in a browser, you probably will not be able to execute individual cells out of order without getting an error. Instead, choose \"Run All\" from the Cell menu after you modify a cell."
16		]
17		},
18		{
19		"cell_type": "code",
20		"execution_count": 25,
21		"metadata": {
22		"collapsed": false,
23		"internals": {
24		"frag_number": 1,
25		"slide_helper": "subslide_end"
26		},
27		"slide_helper": "slide_end",
28		"slideshow": {
29		"slide_type": "fragment"
30		}
31		},
32		"outputs": [],
33		"source": [
34		"import netCDF4 # Note: python is case-sensitive!\n",
35		"import numpy as np"
36		]
37		},
38		{
39		"cell_type": "markdown",
40		"metadata": {
41		"internals": {
42		"frag_helper": "fragment_end",
43		"frag_number": 1,
44		"slide_type": "subslide"
45		},
46		"slideshow": {
47		"slide_type": "slide"
48		}
49		},
50		"source": [
51		"## Opening a file, creating a new Dataset\n",
52		"\n",
53		"Let's create a new, empty netCDF file named 'data/new.nc', opened for writing.\n",
54		"\n",
55		"Be careful, opening a file with 'w' will clobber any existing data (unless `clobber=False` is used, in which case an exception is raised if the file already exists).\n",
56		"\n",
57		"- `mode='r'` is the default.\n",
58		"- `mode='a'` opens an existing file and allows for appending (does not clobber existing data)\n",
59		"- `format` can be one of `NETCDF3_CLASSIC`, `NETCDF3_64BIT`, `NETCDF4_CLASSIC` or `NETCDF4` (default). `NETCDF4_CLASSIC` uses HDF5 for the underlying storage layer (as does `NETCDF4`) but enforces the classic netCDF 3 data model so data can be read with older clients. "
60		]
61		},
62		{
63		"cell_type": "code",
64		"execution_count": 26,
65		"metadata": {
66		"collapsed": false,
67		"internals": {
68		"frag_helper": "fragment_end",
69		"frag_number": 3,
70		"slide_helper": "subslide_end"
71		},
72		"slide_helper": "slide_end",
73		"slideshow": {
74		"slide_type": "fragment"
75		}
76		},
77		"outputs": [
78		{
79		"name": "stdout",
80		"output_type": "stream",
81		"text": [
82		"<type 'netCDF4._netCDF4.Dataset'>\n",
83		"root group (NETCDF4_CLASSIC data model, file format HDF5):\n",
84		" dimensions(sizes): \n",
85		" variables(dimensions): \n",
86		" groups: \n",
87		"\n"
88		]
89		}
90		],
91		"source": [
92		"try: ncfile.close() # just to be safe, make sure dataset is not already open.\n",
93		"except: pass\n",
94		"ncfile = netCDF4.Dataset('data/new.nc',mode='w',format='NETCDF4_CLASSIC') \n",
95		"print(ncfile)"
96		]
97		},
98		{
99		"cell_type": "markdown",
100		"metadata": {
101		"internals": {
102		"frag_helper": "fragment_end",
103		"frag_number": 3,
104		"slide_type": "subslide"
105		},
106		"slideshow": {
107		"slide_type": "slide"
108		}
109		},
110		"source": [
111		"## Creating dimensions\n",
112		"\n",
113		"The ncfile object we created is a container for _dimensions_, _variables_, and _attributes_. First, let's create some dimensions using the [`createDimension`](http://unidata.github.io/netcdf4-python/netCDF4.Dataset-class.html#createDimension) method. \n",
114		"\n",
115		"- Every dimension has a name and a length. \n",
116		"- The name is a string that is used to specify the dimension to be used when creating a variable, and as a key to access the dimension object in the `ncfile.dimensions` dictionary.\n",
117		"\n",
118		"Setting the dimension length to `0` or `None` makes it unlimited, so it can grow. \n",
119		"\n",
120		"- For `NETCDF4` files, any variable's dimension can be unlimited. \n",
121		"- For `NETCDF4_CLASSIC` and `NETCDF3*` files, only one per variable can be unlimited, and it must be the leftmost (fastest varying) dimension."
122		]
123		},
124		{
125		"cell_type": "code",
126		"execution_count": 27,
127		"metadata": {
128		"collapsed": false,
129		"internals": {
130		"frag_helper": "fragment_end",
131		"frag_number": 5,
132		"slide_helper": "subslide_end"
133		},
134		"slide_helper": "slide_end",
135		"slideshow": {
136		"slide_type": "fragment"
137		}
138		},
139		"outputs": [
140		{
141		"name": "stdout",
142		"output_type": "stream",
143		"text": [
144		"('lat', <type 'netCDF4._netCDF4.Dimension'>: name = 'lat', size = 73\n",
145		")\n",
146		"('lon', <type 'netCDF4._netCDF4.Dimension'>: name = 'lon', size = 144\n",
147		")\n",
148		"('time', <type 'netCDF4._netCDF4.Dimension'> (unlimited): name = 'time', size = 0\n",
149		")\n"
150		]
151		}
152		],
153		"source": [
154		"lat_dim = ncfile.createDimension('lat', 73) # latitude axis\n",
155		"lon_dim = ncfile.createDimension('lon', 144) # longitude axis\n",
156		"time_dim = ncfile.createDimension('time', None) # unlimited axis (can be appended to).\n",
157		"for dim in ncfile.dimensions.items():\n",
158		" print(dim)"
159		]
160		},
161		{
162		"cell_type": "markdown",
163		"metadata": {
164		"internals": {
165		"frag_helper": "fragment_end",
166		"frag_number": 5,
167		"slide_type": "subslide"
168		},
169		"slideshow": {
170		"slide_type": "slide"
171		}
172		},
173		"source": [
174		"## Creating attributes\n",
175		"\n",
176		"netCDF attributes can be created just like you would for any python object. \n",
177		"\n",
178		"- Best to adhere to established conventions (like the [CF](http://cfconventions.org/) conventions)\n",
179		"- We won't try to adhere to any specific convention here though."
180		]
181		},
182		{
183		"cell_type": "code",
184		"execution_count": 28,
185		"metadata": {
186		"collapsed": false,
187		"internals": {
188		"frag_helper": "fragment_end",
189		"frag_number": 7
190		},
191		"slideshow": {
192		"slide_type": "fragment"
193		}
194		},
195		"outputs": [
196		{
197		"name": "stdout",
198		"output_type": "stream",
199		"text": [
200		"My model data\n"
201		]
202		}
203		],
204		"source": [
205		"ncfile.title='My model data'\n",
206		"print(ncfile.title)"
207		]
208		},
209		{
210		"cell_type": "markdown",
211		"metadata": {
212		"internals": {
213		"frag_helper": "fragment_end",
214		"frag_number": 8,
215		"slide_helper": "subslide_end"
216		},
217		"slide_helper": "slide_end",
218		"slideshow": {
219		"slide_type": "fragment"
220		}
221		},
222		"source": [
223		"Try adding some more attributes..."
224		]
225		},
226		{
227		"cell_type": "markdown",
228		"metadata": {
229		"internals": {
230		"frag_helper": "fragment_end",
231		"frag_number": 8,
232		"slide_type": "subslide"
233		},
234		"slideshow": {
235		"slide_type": "slide"
236		}
237		},
238		"source": [
239		"## Creating variables\n",
240		"\n",
241		"Now let's add some variables and store some data in them. \n",
242		"\n",
243		"- A variable has a name, a type, a shape, and some data values. \n",
244		"- The shape of a variable is specified by a tuple of dimension names. \n",
245		"- A variable should also have some named attributes, such as 'units', that describe the data.\n",
246		"\n",
247		"The [`createVariable`](http://unidata.github.io/netcdf4-python/netCDF4.Dataset-class.html#createVariable) method takes 3 mandatory args.\n",
248		"\n",
249		"- the 1st argument is the variable name (a string). This is used as the key to access the variable object from the `variables` dictionary.\n",
250		"- the 2nd argument is the datatype (most numpy datatypes supported). \n",
251		"- the third argument is a tuple containing the dimension names (the dimensions must be created first). Unless this is a `NETCDF4` file, any unlimited dimension must be the leftmost one.\n",
252		"- there are lots of optional arguments (many of which are only relevant when `format='NETCDF4'`) to control compression, chunking, fill_value, etc.\n"
253		]
254		},
255		{
256		"cell_type": "code",
257		"execution_count": 29,
258		"metadata": {
259		"collapsed": false,
260		"internals": {
261		"frag_helper": "fragment_end",
262		"frag_number": 10,
263		"slide_helper": "subslide_end"
264		},
265		"slide_helper": "slide_end",
266		"slideshow": {
267		"slide_type": "fragment"
268		}
269		},
270		"outputs": [
271		{
272		"name": "stdout",
273		"output_type": "stream",
274		"text": [
275		"<type 'netCDF4._netCDF4.Variable'>\n",
276		"float64 temp(time, lat, lon)\n",
277		" units: K\n",
278		" standard_name: air_temperature\n",
279		"unlimited dimensions: time\n",
280		"current shape = (0, 73, 144)\n",
281		"filling on, default _FillValue of 9.96920996839e+36 used\n",
282		"\n"
283		]
284		}
285		],
286		"source": [
287		"# Define two variables with the same names as dimensions,\n",
288		"# a conventional way to define \"coordinate variables\".\n",
289		"lat = ncfile.createVariable('lat', np.float32, ('lat',))\n",
290		"lat.units = 'degrees_north'\n",
291		"lat.long_name = 'latitude'\n",
292		"lon = ncfile.createVariable('lon', np.float32, ('lon',))\n",
293		"lon.units = 'degrees_east'\n",
294		"lon.long_name = 'longitude'\n",
295		"time = ncfile.createVariable('time', np.float64, ('time',))\n",
296		"time.units = 'hours since 1800-01-01'\n",
297		"time.long_name = 'time'\n",
298		"# Define a 3D variable to hold the data\n",
299		"temp = ncfile.createVariable('temp',np.float64,('time','lat','lon')) # note: unlimited dimension is leftmost\n",
300		"temp.units = 'K' # degrees Kelvin\n",
301		"temp.standard_name = 'air_temperature' # this is a CF standard name\n",
302		"print(temp)"
303		]
304		},
305		{
306		"cell_type": "markdown",
307		"metadata": {
308		"internals": {
309		"frag_helper": "fragment_end",
310		"frag_number": 10,
311		"slide_type": "subslide"
312		},
313		"slideshow": {
314		"slide_type": "slide"
315		}
316		},
317		"source": [
318		"## Pre-defined variable attributes (read only)\n",
319		"\n",
320		"The netCDF4 module provides some useful pre-defined Python attributes for netCDF variables, such as dimensions, shape, dtype, ndim. \n",
321		"\n",
322		"Note: since no data has been written yet, the length of the 'time' dimension is 0."
323		]
324		},
325		{
326		"cell_type": "code",
327		"execution_count": 30,
328		"metadata": {
329		"collapsed": false,
330		"internals": {
331		"frag_helper": "fragment_end",
332		"frag_number": 12,
333		"slide_helper": "subslide_end"
334		},
335		"slide_helper": "slide_end",
336		"slideshow": {
337		"slide_type": "fragment"
338		}
339		},
340		"outputs": [
341		{
342		"name": "stdout",
343		"output_type": "stream",
344		"text": [
345		"-- Some pre-defined attributes for variable temp:\n",
346		"('temp.dimensions:', (u'time', u'lat', u'lon'))\n",
347		"('temp.shape:', (0, 73, 144))\n",
348		"('temp.dtype:', dtype('float64'))\n",
349		"('temp.ndim:', 3)\n"
350		]
351		}
352		],
353		"source": [
354		"print(\"-- Some pre-defined attributes for variable temp:\")\n",
355		"print(\"temp.dimensions:\", temp.dimensions)\n",
356		"print(\"temp.shape:\", temp.shape)\n",
357		"print(\"temp.dtype:\", temp.dtype)\n",
358		"print(\"temp.ndim:\", temp.ndim)"
359		]
360		},
361		{
362		"cell_type": "markdown",
363		"metadata": {
364		"internals": {
365		"frag_helper": "fragment_end",
366		"frag_number": 12,
367		"slide_type": "subslide"
368		},
369		"slideshow": {
370		"slide_type": "slide"
371		}
372		},
373		"source": [
374		"## Writing data\n",
375		"\n",
376		"To write data to a netCDF variable object, just treat it like a numpy array and assign values to a slice."
377		]
378		},
379		{
380		"cell_type": "code",
381		"execution_count": 31,
382		"metadata": {
383		"collapsed": false,
384		"internals": {
385		"frag_helper": "fragment_end",
386		"frag_number": 14
387		},
388		"slideshow": {
389		"slide_type": "fragment"
390		}
391		},
392		"outputs": [
393		{
394		"name": "stdout",
395		"output_type": "stream",
396		"text": [
397		"('-- Wrote data, temp.shape is now ', (3, 73, 144))\n",
398		"('-- Min/Max values:', 280.00283562143028, 329.99987991477548)\n"
399		]
400		}
401		],
402		"source": [
403		"nlats = len(lat_dim); nlons = len(lon_dim); ntimes = 3\n",
404		"# Write latitudes, longitudes.\n",
405		"# Note: the \":\" is necessary in these \"write\" statements\n",
406		"lat[:] = -90. + (180./nlats)*np.arange(nlats) # south pole to north pole\n",
407		"lon[:] = (180./nlats)*np.arange(nlons) # Greenwich meridian eastward\n",
408		"# create a 3D array of random numbers\n",
409		"data_arr = np.random.uniform(low=280,high=330,size=(ntimes,nlats,nlons))\n",
410		"# Write the data. This writes the whole 3D netCDF variable all at once.\n",
411		"temp[:,:,:] = data_arr # Appends data along unlimited dimension\n",
412		"print(\"-- Wrote data, temp.shape is now \", temp.shape)\n",
413		"# read data back from variable (by slicing it), print min and max\n",
414		"print(\"-- Min/Max values:\", temp[:,:,:].min(), temp[:,:,:].max())"
415		]
416		},
417		{
418		"cell_type": "markdown",
419		"metadata": {
420		"internals": {
421		"frag_helper": "fragment_end",
422		"frag_number": 15,
423		"slide_helper": "subslide_end"
424		},
425		"slide_helper": "slide_end",
426		"slideshow": {
427		"slide_type": "fragment"
428		}
429		},
430		"source": [
431		"- You can just treat a netCDF Variable object like a numpy array and assign values to it.\n",
432		"- Variables automatically grow along unlimited dimensions (unlike numpy arrays)\n",
433		"- The above writes the whole 3D variable all at once, but you can write it a slice at a time instead.\n",
434		"\n",
435		"Let's add another time slice....\n"
436		]
437		},
438		{
439		"cell_type": "code",
440		"execution_count": 32,
441		"metadata": {
442		"collapsed": false,
443		"internals": {
444		"frag_helper": "fragment_end",
445		"frag_number": 15,
446		"slide_type": "subslide"
447		},
448		"slideshow": {
449		"slide_type": "slide"
450		}
451		},
452		"outputs": [
453		{
454		"name": "stdout",
455		"output_type": "stream",
456		"text": [
457		"('-- Wrote more data, temp.shape is now ', (4, 73, 144))\n"
458		]
459		}
460		],
461		"source": [
462		"# create a 2D array of random numbers\n",
463		"data_slice = np.random.uniform(low=280,high=330,size=(nlats,nlons))\n",
464		"temp[3,:,:] = data_slice # Appends the 4th time slice\n",
465		"print(\"-- Wrote more data, temp.shape is now \", temp.shape)"
466		]
467		},
468		{
469		"cell_type": "markdown",
470		"metadata": {
471		"internals": {
472		"frag_helper": "fragment_end",
473		"frag_number": 17
474		},
475		"slideshow": {
476		"slide_type": "fragment"
477		}
478		},
479		"source": [
480		"Note that we have not yet written any data to the time variable. It automatically grew as we appended data along the time dimension to the variable `temp`, but the data is missing."
481		]
482		},
483		{
484		"cell_type": "code",
485		"execution_count": 33,
486		"metadata": {
487		"collapsed": false,
488		"internals": {
489		"frag_helper": "fragment_end",
490		"frag_number": 18,
491		"slide_helper": "subslide_end"
492		},
493		"slide_helper": "slide_end",
494		"slideshow": {
495		"slide_type": "fragment"
496		}
497		},
498		"outputs": [
499		{
500		"name": "stdout",
501		"output_type": "stream",
502		"text": [
503		"<type 'netCDF4._netCDF4.Variable'>\n",
504		"float64 time(time)\n",
505		" units: hours since 1800-01-01\n",
506		" long_name: time\n",
507		"unlimited dimensions: time\n",
508		"current shape = (4,)\n",
509		"filling on, default _FillValue of 9.96920996839e+36 used\n",
510		"\n",
511		"(<class 'numpy.ma.core.MaskedArray'>, masked_array(data = [-- -- -- --],\n",
512		" mask = [ True True True True],\n",
513		" fill_value = 9.96920996839e+36)\n",
514		")\n"
515		]
516		}
517		],
518		"source": [
519		"print(time)\n",
520		"times_arr = time[:]\n",
521		"print(type(times_arr),times_arr) # dashes indicate masked values (where data has not yet been written)"
522		]
523		},
524		{
525		"cell_type": "markdown",
526		"metadata": {
527		"internals": {
528		"frag_helper": "fragment_end",
529		"frag_number": 18,
530		"slide_type": "subslide"
531		},
532		"slideshow": {
533		"slide_type": "slide"
534		}
535		},
536		"source": [
537		"Let's add write some data into the time variable. \n",
538		"\n",
539		"- Given a set of datetime instances, use date2num to convert to numeric time values and then write that data to the variable."
540		]
541		},
542		{
543		"cell_type": "code",
544		"execution_count": 34,
545		"metadata": {
546		"collapsed": false,
547		"internals": {
548		"frag_helper": "fragment_end",
549		"frag_number": 20,
550		"slide_helper": "subslide_end"
551		},
552		"slide_helper": "slide_end",
553		"slideshow": {
554		"slide_type": "fragment"
555		}
556		},
557		"outputs": [
558		{
559		"name": "stdout",
560		"output_type": "stream",
561		"text": [
562		"[datetime.datetime(2014, 10, 1, 0, 0), datetime.datetime(2014, 10, 2, 0, 0), datetime.datetime(2014, 10, 3, 0, 0), datetime.datetime(2014, 10, 4, 0, 0)]\n",
563		"(array([ 1882440., 1882464., 1882488., 1882512.]), u'hours since 1800-01-01')\n",
564		"[datetime.datetime(2014, 10, 1, 0, 0) datetime.datetime(2014, 10, 2, 0, 0)\n",
565		" datetime.datetime(2014, 10, 3, 0, 0) datetime.datetime(2014, 10, 4, 0, 0)]\n"
566		]
567		}
568		],
569		"source": [
570		"from datetime import datetime\n",
571		"from netCDF4 import date2num,num2date\n",
572		"# 1st 4 days of October.\n",
573		"dates = [datetime(2014,10,1,0),datetime(2014,10,2,0),datetime(2014,10,3,0),datetime(2014,10,4,0)]\n",
574		"print(dates)\n",
575		"times = date2num(dates, time.units)\n",
576		"print(times, time.units) # numeric values\n",
577		"time[:] = times\n",
578		"# read time data back, convert to datetime instances, check values.\n",
579		"print(num2date(time[:],time.units))"
580		]
581		},
582		{
583		"cell_type": "markdown",
584		"metadata": {
585		"internals": {
586		"frag_helper": "fragment_end",
587		"frag_number": 20,
588		"slide_type": "subslide"
589		},
590		"slideshow": {
591		"slide_type": "slide"
592		}
593		},
594		"source": [
595		"## Closing a netCDF file\n",
596		"\n",
597		"It's important to close a netCDF file you opened for writing:\n",
598		"\n",
599		"- flushes buffers to make sure all data gets written\n",
600		"- releases memory resources used by open netCDF files"
601		]
602		},
603		{
604		"cell_type": "code",
605		"execution_count": 35,
606		"metadata": {
607		"collapsed": false,
608		"internals": {
609		"frag_helper": "fragment_end",
610		"frag_number": 22,
611		"slide_helper": "subslide_end"
612		},
613		"slide_helper": "slide_end",
614		"slideshow": {
615		"slide_type": "fragment"
616		}
617		},
618		"outputs": [
619		{
620		"name": "stdout",
621		"output_type": "stream",
622		"text": [
623		"<type 'netCDF4._netCDF4.Dataset'>\n",
624		"root group (NETCDF4_CLASSIC data model, file format HDF5):\n",
625		" title: My model data\n",
626		" dimensions(sizes): lat(73), lon(144), time(4)\n",
627		" variables(dimensions): float32 \u001b[4mlat\u001b[0m(lat), float32 \u001b[4mlon\u001b[0m(lon), float64 \u001b[4mtime\u001b[0m(time), float64 \u001b[4mtemp\u001b[0m(time,lat,lon)\n",
628		" groups: \n",
629		"\n",
630		"Dataset is closed!\n"
631		]
632		}
633		],
634		"source": [
635		"# first print the Dataset object to see what we've got\n",
636		"print(ncfile)\n",
637		"# close the Dataset.\n",
638		"ncfile.close(); print('Dataset is closed!')"
639		]
640		},
641		{
642		"cell_type": "markdown",
643		"metadata": {
644		"internals": {
645		"frag_helper": "fragment_end",
646		"frag_number": 22,
647		"slide_type": "subslide"
648		},
649		"slideshow": {
650		"slide_type": "slide"
651		}
652		},
653		"source": [
654		"# Advanced features\n",
655		"\n",
656		"So far we've only exercised features associated with the old netCDF version 3 data model. netCDF version 4 adds a lot of new functionality that comes with the more flexible HDF5 storage layer. \n",
657		"\n",
658		"Let's create a new file with `format='NETCDF4'` so we can try out some of these features."
659		]
660		},
661		{
662		"cell_type": "code",
663		"execution_count": 36,
664		"metadata": {
665		"collapsed": false,
666		"internals": {
667		"frag_helper": "fragment_end",
668		"frag_number": 25,
669		"slide_helper": "subslide_end"
670		},
671		"slide_helper": "slide_end",
672		"slideshow": {
673		"slide_type": "fragment"
674		}
675		},
676		"outputs": [
677		{
678		"name": "stdout",
679		"output_type": "stream",
680		"text": [
681		"<type 'netCDF4._netCDF4.Dataset'>\n",
682		"root group (NETCDF4 data model, file format HDF5):\n",
683		" dimensions(sizes): \n",
684		" variables(dimensions): \n",
685		" groups: \n",
686		"\n"
687		]
688		}
689		],
690		"source": [
691		"ncfile = netCDF4.Dataset('data/new2.nc','w',format='NETCDF4')\n",
692		"print(ncfile)"
693		]
694		},
695		{
696		"cell_type": "markdown",
697		"metadata": {
698		"internals": {
699		"frag_helper": "fragment_end",
700		"frag_number": 25,
701		"slide_type": "subslide"
702		},
703		"slideshow": {
704		"slide_type": "slide"
705		}
706		},
707		"source": [
708		"## Creating Groups\n",
709		"\n",
710		"netCDF version 4 added support for organizing data in hierarchical groups.\n",
711		"\n",
712		"- analagous to directories in a filesystem. \n",
713		"- Groups serve as containers for variables, dimensions and attributes, as well as other groups. \n",
714		"- A `netCDF4.Dataset` creates a special group, called the 'root group', which is similar to the root directory in a unix filesystem. \n",
715		"\n",
716		"- groups are created using the [`createGroup`](http://unidata.github.io/netcdf4-python/netCDF4.Dataset-class.html#createGroup) method.\n",
717		"- takes a single argument (a string, which is the name of the Group instance). This string is used as a key to access the group instances in the `groups` dictionary.\n",
718		"\n",
719		"Here we create two groups to hold data for two different model runs."
720		]
721		},
722		{
723		"cell_type": "code",
724		"execution_count": 37,
725		"metadata": {
726		"collapsed": false,
727		"internals": {
728		"frag_helper": "fragment_end",
729		"frag_number": 27,
730		"slide_helper": "subslide_end"
731		},
732		"slide_helper": "slide_end",
733		"slideshow": {
734		"slide_type": "fragment"
735		}
736		},
737		"outputs": [
738		{
739		"name": "stdout",
740		"output_type": "stream",
741		"text": [
742		"('model_run1', <type 'netCDF4._netCDF4.Group'>\n",
743		"group /model_run1:\n",
744		" dimensions(sizes): \n",
745		" variables(dimensions): \n",
746		" groups: \n",
747		")\n",
748		"('model_run2', <type 'netCDF4._netCDF4.Group'>\n",
749		"group /model_run2:\n",
750		" dimensions(sizes): \n",
751		" variables(dimensions): \n",
752		" groups: \n",
753		")\n"
754		]
755		}
756		],
757		"source": [
758		"grp1 = ncfile.createGroup('model_run1')\n",
759		"grp2 = ncfile.createGroup('model_run2')\n",
760		"for grp in ncfile.groups.items():\n",
761		" print(grp)"
762		]
763		},
764		{
765		"cell_type": "markdown",
766		"metadata": {
767		"internals": {
768		"frag_helper": "fragment_end",
769		"frag_number": 27,
770		"slide_type": "subslide"
771		},
772		"slideshow": {
773		"slide_type": "slide"
774		}
775		},
776		"source": [
777		"Create some dimensions in the root group."
778		]
779		},
780		{
781		"cell_type": "code",
782		"execution_count": 38,
783		"metadata": {
784		"collapsed": false,
785		"internals": {
786		"frag_helper": "fragment_end",
787		"frag_number": 29
788		},
789		"slideshow": {
790		"slide_type": "fragment"
791		}
792		},
793		"outputs": [],
794		"source": [
795		"lat_dim = ncfile.createDimension('lat', 73) # latitude axis\n",
796		"lon_dim = ncfile.createDimension('lon', 144) # longitude axis\n",
797		"time_dim = ncfile.createDimension('time', None) # unlimited axis (can be appended to)."
798		]
799		},
800		{
801		"cell_type": "markdown",
802		"metadata": {
803		"internals": {
804		"frag_helper": "fragment_end",
805		"frag_number": 30
806		},
807		"slideshow": {
808		"slide_type": "fragment"
809		}
810		},
811		"source": [
812		"Now create a variable in grp1 and grp2. The library will search recursively upwards in the group tree to find the dimensions (which in this case are defined one level up).\n",
813		"\n",
814		"- These variables are create with zlib compression, another nifty feature of netCDF 4. \n",
815		"- The data are automatically compressed when data is written to the file, and uncompressed when the data is read. \n",
816		"- This can really save disk space, especially when used in conjunction with the [least_significant_digit](http://unidata.github.io/netcdf4-python/netCDF4.Dataset-class.html#createVariable) keyword argument, which causes the data to be quantized (truncated) before compression. This makes the compression lossy, but more efficient."
817		]
818		},
819		{
820		"cell_type": "code",
821		"execution_count": 39,
822		"metadata": {
823		"collapsed": false,
824		"internals": {
825		"frag_helper": "fragment_end",
826		"frag_number": 31,
827		"slide_helper": "subslide_end"
828		},
829		"slide_helper": "slide_end",
830		"slideshow": {
831		"slide_type": "fragment"
832		}
833		},
834		"outputs": [
835		{
836		"name": "stdout",
837		"output_type": "stream",
838		"text": [
839		"('model_run1', <type 'netCDF4._netCDF4.Group'>\n",
840		"group /model_run1:\n",
841		" dimensions(sizes): \n",
842		" variables(dimensions): float64 \u001b[4mtemp\u001b[0m(time,lat,lon)\n",
843		" groups: \n",
844		")\n",
845		"('model_run2', <type 'netCDF4._netCDF4.Group'>\n",
846		"group /model_run2:\n",
847		" dimensions(sizes): \n",
848		" variables(dimensions): float64 \u001b[4mtemp\u001b[0m(time,lat,lon)\n",
849		" groups: \n",
850		")\n"
851		]
852		}
853		],
854		"source": [
855		"temp1 = grp1.createVariable('temp',np.float64,('time','lat','lon'),zlib=True)\n",
856		"temp2 = grp2.createVariable('temp',np.float64,('time','lat','lon'),zlib=True)\n",
857		"for grp in ncfile.groups.items(): # shows that each group now contains 1 variable\n",
858		" print(grp)"
859		]
860		},
861		{
862		"cell_type": "markdown",
863		"metadata": {
864		"internals": {
865		"frag_helper": "fragment_end",
866		"frag_number": 31,
867		"slide_type": "subslide"
868		},
869		"slideshow": {
870		"slide_type": "slide"
871		}
872		},
873		"source": [
874		"##Creating a variable with a compound data type\n",
875		"\n",
876		"- Compound data types map directly to numpy structured (a.k.a 'record' arrays). \n",
877		"- Structured arrays are akin to C structs, or derived types in Fortran. \n",
878		"- They allow for the construction of table-like structures composed of combinations of other data types, including other compound types. \n",
879		"- Might be useful for representing multiple parameter values at each point on a grid, or at each time and space location for scattered (point) data. \n",
880		"\n",
881		"Here we create a variable with a compound data type to represent complex data (there is no native complex data type in netCDF). \n",
882		"\n",
883		"- The compound data type is created with the [`createCompoundType`](http://unidata.github.io/netcdf4-python/netCDF4.Dataset-class.html#createCompoundType) method."
884		]
885		},
886		{
887		"cell_type": "code",
888		"execution_count": 40,
889		"metadata": {
890		"collapsed": false,
891		"internals": {
892		"frag_helper": "fragment_end",
893		"frag_number": 33,
894		"slide_helper": "subslide_end"
895		},
896		"slide_helper": "slide_end",
897		"slideshow": {
898		"slide_type": "fragment"
899		}
900		},
901		"outputs": [
902		{
903		"name": "stdout",
904		"output_type": "stream",
905		"text": [
906		"<type 'netCDF4._netCDF4.Variable'>\n",
907		"compound cmplx_var(time, lat, lon)\n",
908		"compound data type: [('real', '<f8'), ('imag', '<f8')]\n",
909		"path = /model_run1\n",
910		"unlimited dimensions: time\n",
911		"current shape = (1, 73, 144)\n",
912		"\n",
913		"(dtype([('real', '<f8'), ('imag', '<f8')]), (73, 144), (0.578177705604801, 0.18086070805676357))\n"
914		]
915		}
916		],
917		"source": [
918		"# create complex128 numpy structured data type\n",
919		"complex128 = np.dtype([('real',np.float64),('imag',np.float64)])\n",
920		"# using this numpy dtype, create a netCDF compound data type object\n",
921		"# the string name can be used as a key to access the datatype from the cmptypes dictionary.\n",
922		"complex128_t = ncfile.createCompoundType(complex128,'complex128')\n",
923		"# create a variable with this data type, write some data to it.\n",
924		"cmplxvar = grp1.createVariable('cmplx_var',complex128_t,('time','lat','lon'))\n",
925		"# write some data to this variable\n",
926		"# first create some complex random data\n",
927		"nlats = len(lat_dim); nlons = len(lon_dim)\n",
928		"data_arr_cmplx = np.random.uniform(size=(nlats,nlons))+1.j*np.random.uniform(size=(nlats,nlons))\n",
929		"# write this complex data to a numpy complex128 structured array\n",
930		"data_arr = np.empty((nlats,nlons),complex128)\n",
931		"data_arr['real'] = data_arr_cmplx.real; data_arr['imag'] = data_arr_cmplx.imag\n",
932		"cmplxvar[0] = data_arr # write the data to the variable (appending to time dimension)\n",
933		"print(cmplxvar)\n",
934		"data_out = cmplxvar[0] # read one value of data back from variable\n",
935		"print(data_out.dtype, data_out.shape, data_out[0,0])"
936		]
937		},
938		{
939		"cell_type": "markdown",
940		"metadata": {
941		"internals": {
942		"frag_helper": "fragment_end",
943		"frag_number": 33,
944		"slide_type": "subslide"
945		},
946		"slideshow": {
947		"slide_type": "slide"
948		}
949		},
950		"source": [
951		"##Creating a variable with a variable-length (vlen) data type\n",
952		"\n",
953		"netCDF 4 has support for variable-length or \"ragged\" arrays. These are arrays of variable length sequences having the same type. \n",
954		"\n",
955		"- To create a variable-length data type, use the [`createVLType`](http://unidata.github.io/netcdf4-python/netCDF4.Dataset-class.html#createVLType) method.\n",
956		"- The numpy datatype of the variable-length sequences and the name of the new datatype must be specified. "
957		]
958		},
959		{
960		"cell_type": "code",
961		"execution_count": 41,
962		"metadata": {
963		"collapsed": false,
964		"internals": {
965		"frag_helper": "fragment_end",
966		"frag_number": 35
967		},
968		"slideshow": {
969		"slide_type": "fragment"
970		}
971		},
972		"outputs": [],
973		"source": [
974		"vlen_t = ncfile.createVLType(np.int64, 'phony_vlen')"
975		]
976		},
977		{
978		"cell_type": "markdown",
979		"metadata": {
980		"internals": {
981		"frag_helper": "fragment_end",
982		"frag_number": 36
983		},
984		"slideshow": {
985		"slide_type": "fragment"
986		}
987		},
988		"source": [
989		"A new variable can then be created using this datatype."
990		]
991		},
992		{
993		"cell_type": "code",
994		"execution_count": 42,
995		"metadata": {
996		"collapsed": false,
997		"internals": {
998		"frag_helper": "fragment_end",
999		"frag_number": 37,
1000		"slide_helper": "subslide_end"
1001		},
1002		"slide_helper": "slide_end",
1003		"slideshow": {
1004		"slide_type": "fragment"
1005		}
1006		},
1007		"outputs": [],
1008		"source": [
1009		"vlvar = grp2.createVariable('phony_vlen_var', vlen_t, ('time','lat','lon'))"
1010		]
1011		},
1012		{
1013		"cell_type": "markdown",
1014		"metadata": {
1015		"internals": {
1016		"frag_helper": "fragment_end",
1017		"frag_number": 37,
1018		"slide_type": "subslide"
1019		},
1020		"slideshow": {
1021		"slide_type": "slide"
1022		}
1023		},
1024		"source": [
1025		"Since there is no native vlen datatype in numpy, vlen arrays are represented in python as object arrays (arrays of dtype `object`). \n",
1026		"\n",
1027		"- These are arrays whose elements are Python object pointers, and can contain any type of python object. \n",
1028		"- For this application, they must contain 1-D numpy arrays all of the same type but of varying length. \n",
1029		"- Fill with 1-D random numpy int64 arrays of random length between 1 and 10."
1030		]
1031		},
1032		{
1033		"cell_type": "code",
1034		"execution_count": 43,
1035		"metadata": {
1036		"collapsed": false,
1037		"internals": {
1038		"frag_helper": "fragment_end",
1039		"frag_number": 39,
1040		"slide_helper": "subslide_end"
1041		},
1042		"slide_helper": "slide_end",
1043		"slideshow": {
1044		"slide_type": "fragment"
1045		}
1046		},
1047		"outputs": [
1048		{
1049		"name": "stdout",
1050		"output_type": "stream",
1051		"text": [
1052		"<type 'netCDF4._netCDF4.Variable'>\n",
1053		"vlen phony_vlen_var(time, lat, lon)\n",
1054		"vlen data type: int64\n",
1055		"path = /model_run2\n",
1056		"unlimited dimensions: time\n",
1057		"current shape = (1, 73, 144)\n",
1058		"\n",
1059		"('data =\\n', array([[[array([0, 4, 0, 9, 2, 2, 2, 4, 2]), array([7, 5, 4, 4, 9, 8, 0]),\n",
1060		" array([3, 6, 6, 8, 2, 7]), ..., array([5, 0, 0, 8, 8, 1, 5, 3]),\n",
1061		" array([4, 2, 7]), array([0])],\n",
1062		" [array([5, 6, 6, 6, 1, 0, 7]), array([7]),\n",
1063		" array([7, 5, 8, 9, 6, 9, 3]), ..., array([0, 6, 5, 4]),\n",
1064		" array([7, 1, 9, 7, 7, 2]), array([1, 4, 0])],\n",
1065		" [array([4, 3, 1]), array([6, 3, 9, 7, 8]), array([8]), ...,\n",
1066		" array([6, 5, 8, 0]), array([0]), array([0, 9, 6, 2, 4])],\n",
1067		" ..., \n",
1068		" [array([8, 4, 4]), array([4, 1, 6]), array([1, 4, 2, 3, 9]), ...,\n",
1069		" array([9, 1]), array([7, 2, 5, 1, 5, 8, 2]),\n",
1070		" array([2, 9, 9, 1, 4, 6, 3, 5, 2])],\n",
1071		" [array([4, 7, 9, 8, 2, 3, 6, 6]),\n",
1072		" array([1, 4, 1, 6, 1, 1, 2, 3, 9]),\n",
1073		" array([9, 5, 6, 2, 4, 3, 8, 2, 9]), ..., array([9, 5, 7]),\n",
1074		" array([3, 9]), array([4, 2, 6, 9])],\n",
1075		" [array([8, 9, 9, 2, 2, 8, 8, 5]), array([3]),\n",
1076		" array([8, 8, 0, 2, 9, 2, 3, 0, 9]), ..., array([7]),\n",
1077		" array([5, 1, 0, 6, 8, 6]), array([8, 6, 3, 6, 9, 8, 4, 2, 5])]]], dtype=object))\n"
1078		]
1079		}
1080		],
1081		"source": [
1082		"vlen_data = np.empty((nlats,nlons),object)\n",
1083		"for i in range(nlons):\n",
1084		" for j in range(nlats):\n",
1085		" size = np.random.randint(1,10,size=1) # random length of sequence\n",
1086		" vlen_data[j,i] = np.random.randint(0,10,size=size)# generate random sequence\n",
1087		"vlvar[0] = vlen_data # append along unlimited dimension (time)\n",
1088		"print(vlvar)\n",
1089		"print('data =\\n',vlvar[:])"
1090		]
1091		},
1092		{
1093		"cell_type": "markdown",
1094		"metadata": {
1095		"internals": {
1096		"frag_helper": "fragment_end",
1097		"frag_number": 39,
1098		"slide_type": "subslide"
1099		},
1100		"slideshow": {
1101		"slide_type": "slide"
1102		}
1103		},
1104		"source": [
1105		"Close the Dataset and examine the contents with ncdump."
1106		]
1107		},
1108		{
1109		"cell_type": "code",
1110		"execution_count": 44,
1111		"metadata": {
1112		"collapsed": false,
1113		"internals": {
1114		"frag_helper": "fragment_end",
1115		"frag_number": 41,
1116		"slide_helper": "subslide_end"
1117		},
1118		"slide_helper": "slide_end",
1119		"slideshow": {
1120		"slide_type": "fragment"
1121		}
1122		},
1123		"outputs": [
1124		{
1125		"name": "stdout",
1126		"output_type": "stream",
1127		"text": [
1128		"netcdf new2 {\r\n",
1129		"types:\r\n",
1130		" compound complex128 {\r\n",
1131		" double real ;\r\n",
1132		" double imag ;\r\n",
1133		" }; // complex128\r\n",
1134		" int64(*) phony_vlen ;\r\n",
1135		"dimensions:\r\n",
1136		"\tlat = 73 ;\r\n",
1137		"\tlon = 144 ;\r\n",
1138		"\ttime = UNLIMITED ; // (1 currently)\r\n",
1139		"\r\n",
1140		"group: model_run1 {\r\n",
1141		" variables:\r\n",
1142		" \tdouble temp(time, lat, lon) ;\r\n",
1143		" \tcomplex128 cmplx_var(time, lat, lon) ;\r\n",
1144		" } // group model_run1\r\n",
1145		"\r\n",
1146		"group: model_run2 {\r\n",
1147		" variables:\r\n",
1148		" \tdouble temp(time, lat, lon) ;\r\n",
1149		" \tphony_vlen phony_vlen_var(time, lat, lon) ;\r\n",
1150		" } // group model_run2\r\n",
1151		"}\r\n"
1152		]
1153		}
1154		],
1155		"source": [
1156		"ncfile.close()\n",
1157		"!ncdump -h data/new2.nc"
1158		]
1159		},
1160		{
1161		"cell_type": "markdown",
1162		"metadata": {
1163		"internals": {
1164		"frag_helper": "fragment_end",
1165		"frag_number": 41,
1166		"slide_helper": "subslide_end",
1167		"slide_type": "subslide"
1168		},
1169		"slide_helper": "slide_end",
1170		"slideshow": {
1171		"slide_type": "slide"
1172		}
1173		},
1174		"source": [
1175		"##Other interesting and useful projects using netcdf4-python\n",
1176		"\n",
1177		"- [xarray](https://xarray.pydata.org/en/stable/): N-dimensional variant of the core [pandas](https://pandas.pydata.org) data structure that can operate on netcdf variables.\n",
1178		"- [Iris](https://scitools.org.uk/iris/docs/latest/): a data model to create a data abstraction layer which isolates analysis and visualisation code from data format specifics. Uses netcdf4-python to access netcdf data (can also handle GRIB).\n",
1179		"- [Dask](https://dask.org/): Virtual large arrays (from netcdf variables) with lazy evaluation.\n",
1180		"- [cf-python](https://cfpython.bitbucket.io/): Implements the [CF](http://cfconventions.org) data model for the reading, writing and processing of data and metadata. "
1181		]
1182		}
1183		],
1184		"metadata": {
1185		"kernelspec": {
1186		"display_name": "Python 2",
1187		"language": "python",
1188		"name": "python2"
1189		},
1190		"language_info": {
1191		"codemirror_mode": {
1192		"name": "ipython",
1193		"version": 2
1194		},
1195		"file_extension": ".py",
1196		"mimetype": "text/x-python",
1197		"name": "python",
1198		"nbconvert_exporter": "python",
1199		"pygments_lexer": "ipython2",
1200		"version": "2.7.9"
1201		}
1202		},
1203		"nbformat": 4,
1204		"nbformat_minor": 0
1205		}

+14

-54

setup.cfg less more

0		# Rename this file to setup.cfg to set build options.
1		# Follow instructions below for editing.
2	0	[options]
3		# if true, the nc-config script (installed with netcdf 4.1.2 and higher)
4		# will be used to determine the locations of required libraries.
5		# Usually, nothing else is needed.
6		use_ncconfig=True
7		# path to nc-config script (use if not found in unix PATH).
8		#ncconfig=/usr/local/bin/nc-config
	1	use_ncconfig = True
	2
9	3	[directories]
10		#
11		# If nc-config doesn't do the trick, you can specify the locations
12		# of the libraries and headers manually below
13		#
14		# uncomment and set to netCDF install location.
15		# Include files should be located in netCDF4_dir/include and
16		# the library should be located in netCDF4_dir/lib.
17		# If the libraries and include files are installed in separate locations,
18		# use netCDF4_libdir and netCDF4_incdir to specify the locations
19		# separately.
20		#netCDF4_dir = /usr/local
21		# uncomment and set to HDF5 install location.
22		# Include files should be located in HDF5_dir/include and
23		# the library should be located in HDF5_dir/lib.
24		# If the libraries and include files are installed in separate locations,
25		# use HDF5_libdir and HDF5_incdir to specify the locations
26		# separately.
27		#HDF5_dir = /usr/local
28		# if HDF5 was built with szip support as a static lib,
29		# uncomment and set to szip lib install location.
30		# If the libraries and include files are installed in separate locations,
31		# use szip_libdir and szip_incdir.
32		#szip_dir = /usr/local
33		# if netcdf lib was build statically with HDF4 support,
34		# uncomment and set to hdf4 lib (libmfhdf and libdf) nstall location.
35		# If the libraries and include files are installed in separate locations,
36		# use hdf4_libdir and hdf4_incdir.
37		#hdf4_dir = /usr/local
38		# if netcdf lib was build statically with HDF4 support,
39		# uncomment and set to jpeg lib install location (hdf4 needs jpeg).
40		# If the libraries and include files are installed in separate locations,
41		# use jpeg_libdir and jpeg_incdir.
42		#jpeg_dir = /usr/local
43		# if netcdf lib was build statically with OpenDAP support,
44		# uncomment and set to curl lib install location.
45		# If the libraries and include files are installed in separate locations,
46		# use curl_libdir and curl_incdir.
47		#curl_dir = /usr/local
48		# location of mpi.h (needed for parallel support)
49		#mpi_incdir=/opt/local/include/mpich-mp
	4
50	5	[check-manifest]
51		ignore =
52		.gitignore
53		README.gh-pages
54		README.release
55		examples/data/*nc
56		examples/*ipynb
	6	ignore =
	7	.gitignore
	8	README.gh-pages
	9	README.release
	10	examples/data/*nc
	11	examples/*ipynb
	12
	13	[egg_info]
	14	tag_build =
	15	tag_date = 0
	16

+322

-335

src/netCDF4/_netCDF4.pyx less more

0	0	"""
1		Version 1.5.6
	1	Version 1.5.7
2	2	-------------
3	3
4	4	# Introduction

22	22	compound types containing enums, or vlens containing compound
23	23	types) are not supported.
24	24
25		## Download
26
27		- Latest bleeding-edge code from the
	25	## Quick Install
	26
	27	- the easiest way to get going is to install via `pip install netCDF4`.
	28	(or if you use the [conda](http://conda.io) package manager `conda install -c conda_forge netCDF4`).
	29
	30	## Developer Install
	31
	32	- Clone the
28	33	[github repository](http://github.com/Unidata/netcdf4-python).
29		- Latest [releases](https://pypi.python.org/pypi/netCDF4)
30		(source code and binary installers).
31
32		## Requires
33
34		- Python 3.6 or later.
35		- [numpy array module](http://numpy.scipy.org), version 1.10.0 or later.
36		- [Cython](http://cython.org), version 0.21 or later.
37		- [setuptools](https://pypi.python.org/pypi/setuptools), version 18.0 or
38		later.
39		- The HDF5 C library version 1.8.4-patch1 or higher
40		from [](ftp://ftp.hdfgroup.org/HDF5/current/src).
41		***netCDF version 4.4.1 or higher is recommended if using HDF5 1.10.x -
42		otherwise resulting files may be unreadable by clients using earlier
43		versions of HDF5. For netCDF < 4.4.1, HDF5 version 1.8.x is recommended.***
44		Be sure to build with `--enable-hl --enable-shared`.
45		- [Libcurl](http://curl.haxx.se/libcurl), if you want
46		[OPeNDAP](http://opendap.org) support.
47		- [HDF4](http://www.hdfgroup.org/products/hdf4), if you want
48		to be able to read HDF4 "Scientific Dataset" (SD) files.
49		- The netCDF-4 C library from the [github releases
50		page](https://github.com/Unidata/netcdf-c/releases).
51		Version 4.1.1 or higher is required (4.2 or higher recommended).
52		Be sure to build with `--enable-netcdf-4 --enable-shared`, and set
53		`CPPFLAGS="-I $HDF5_DIR/include"` and `LDFLAGS="-L $HDF5_DIR/lib"`,
54		where `$HDF5_DIR` is the directory where HDF5 was installed.
55		If you want [OPeNDAP](http://opendap.org) support, add `--enable-dap`.
56		If you want HDF4 SD support, add `--enable-hdf4` and add
57		the location of the HDF4 headers and library to `$CPPFLAGS` and `$LDFLAGS`.
58		- for MPI parallel IO support, an MPI-enabled versions of the netcdf library
59		is required, as is the [mpi4py](http://mpi4py.scipy.org) python module.
60		Parallel IO further depends on the existence of MPI-enabled HDF5 or the
61		[PnetCDF](https://parallel-netcdf.github.io/) library.
62		- [cftime](https://github.com/Unidata/cftime) for
63		time and date handling utility functions.
64
65
66		## Install
67
68		- install the requisite python modules and C libraries (see above). It's
69		easiest if all the C libs are built as shared libraries.
70		- By default, the utility `nc-config`, installed with netcdf 4.1.2 or higher,
71		will be run used to determine where all the dependencies live.
	34	- Make sure the dependencies are satisfied (Python 3.6 or later,
	35	[numpy](http://numpy.scipy.org),
	36	[Cython](http://cython.org),
	37	[cftime](https://github.com/Unidata/cftime),
	38	[setuptools](https://pypi.python.org/pypi/setuptools),
	39	the [HDF5 C library](https://www.hdfgroup.org/solutions/hdf5/),
	40	and the [netCDF C library](https://www.unidata.ucar.edu/software/netcdf/)).
	41	For MPI parallel IO support, an MPI-enabled versions of the netcdf library
	42	is required, as is [mpi4py](http://mpi4py.scipy.org).
	43	Parallel IO further depends on the existence of MPI-enabled HDF5 or the
	44	[PnetCDF](https://parallel-netcdf.github.io/) library.
	45	- By default, the utility `nc-config` (installed with netcdf-c)
	46	will be run used to determine where all the dependencies live.
72	47	- If `nc-config` is not in your default `PATH`, you can set the `NETCDF4_DIR`
73		environment variable and `setup.py` will look in `$NETCDF4_DIR/bin`.
74		You can also use the file `setup.cfg` to set the path to `nc-config`, or
75		enter the paths to the libraries and include files manually. Just edit the `setup.cfg` file
76		in a text editor and follow the instructions in the comments.
77		To disable the use of `nc-config`, set the env var `USE_NCCONFIG` to 0.
78		To disable the use of `setup.cfg`, set `USE_SETUPCFG` to 0.
79		As a last resort, the library and include paths can be set via environment variables.
80		If you go this route, set `USE_NCCONFIG` and `USE_SETUPCFG` to 0, and specify
81		`NETCDF4_LIBDIR`, `NETCDF4_INCDIR`, `HDF5_LIBDIR` and `HDF5_INCDIR`.
82		Similarly, environment variables
83		(all capitalized) can be used to set the include and library paths for
84		`hdf4`, `szip`, `jpeg`, `curl` and `zlib`. If the dependencies are not found
85		in any of the paths specified by environment variables, then standard locations
86		(such as `/usr` and `/usr/local`) are searched.
	48	environment variable and `setup.py` will look in `$NETCDF4_DIR/bin`.
	49	You can also use the file `setup.cfg` to set the path to `nc-config`, or
	50	enter the paths to the libraries and include files manually. Just edit the `setup.cfg` file
	51	in a text editor and follow the instructions in the comments.
	52	To disable the use of `nc-config`, set the env var `USE_NCCONFIG` to 0.
	53	To disable the use of `setup.cfg`, set `USE_SETUPCFG` to 0.
	54	As a last resort, the library and include paths can be set via environment variables.
	55	If you go this route, set `USE_NCCONFIG` and `USE_SETUPCFG` to 0, and specify
	56	`NETCDF4_LIBDIR`, `NETCDF4_INCDIR`, `HDF5_LIBDIR` and `HDF5_INCDIR`.
	57	If the dependencies are not found
	58	in any of the paths specified by environment variables, then standard locations
	59	(such as `/usr` and `/usr/local`) are searched.
87	60	- run `python setup.py build`, then `python setup.py install` (as root if
88		necessary). `pip install` can be used to install pre-compiled binary wheels from
89		[pypi](https://pypi.org/project/netCDF4).
	61	necessary).
90	62	- run the tests in the 'test' directory by running `python run_all.py`.
91	63
92	64	# Tutorial

109	81
110	82	## Creating/Opening/Closing a netCDF file
111	83
112		To create a netCDF file from python, you simply call the [Dataset](#Dataset)
	84	To create a netCDF file from python, you simply call the `Dataset`
113	85	constructor. This is also the method used to open an existing netCDF
114	86	file. If the file is open for write access (`mode='w', 'r+'` or `'a'`), you may
115	87	write any type of data including new dimensions, groups, variables and

133	105	keyword in the `Dataset` constructor. The default format is
134	106	`NETCDF4`. To see how a given file is formatted, you can examine the
135	107	`data_model` attribute. Closing the netCDF file is
136		accomplished via the [Dataset.close](#Dataset.close) method of the [Dataset](#Dataset)
	108	accomplished via the `Dataset.close` method of the `Dataset`
137	109	instance.
138	110
139	111	Here's an example:

147	119	```
148	120
149	121	Remote [OPeNDAP](http://opendap.org)-hosted datasets can be accessed for
150		reading over http if a URL is provided to the [Dataset](#Dataset) constructor instead of a
	122	reading over http if a URL is provided to the `Dataset` constructor instead of a
151	123	filename. However, this requires that the netCDF library be built with
152	124	OPenDAP support, via the `--enable-dap` configure option (added in
153	125	version 4.0.1).

158	130	netCDF version 4 added support for organizing data in hierarchical
159	131	groups, which are analogous to directories in a filesystem. Groups serve
160	132	as containers for variables, dimensions and attributes, as well as other
161		groups. A [Dataset](#Dataset) creates a special group, called
	133	groups. A `Dataset` creates a special group, called
162	134	the 'root group', which is similar to the root directory in a unix
163		filesystem. To create [Group](#Group) instances, use the
164		[Dataset.createGroup](#Dataset.createGroup) method of a [Dataset](#Dataset) or [Group](#Group)
165		instance. [Dataset.createGroup](#Dataset.createGroup) takes a single argument, a
166		python string containing the name of the new group. The new [Group](#Group)
	135	filesystem. To create `Group` instances, use the
	136	`Dataset.createGroup` method of a `Dataset` or `Group`
	137	instance. `Dataset.createGroup` takes a single argument, a
	138	python string containing the name of the new group. The new `Group`
167	139	instances contained within the root group can be accessed by name using
168		the `groups` dictionary attribute of the [Dataset](#Dataset) instance. Only
	140	the `groups` dictionary attribute of the `Dataset` instance. Only
169	141	`NETCDF4` formatted files support Groups, if you try to create a Group
170	142	in a netCDF 3 file you will get an error message.
171	143

186	158	>>>
187	159	```
188	160
189		Groups can exist within groups in a [Dataset](#Dataset), just as directories
190		exist within directories in a unix filesystem. Each [Group](#Group) instance
	161	Groups can exist within groups in a `Dataset`, just as directories
	162	exist within directories in a unix filesystem. Each `Group` instance
191	163	has a `groups` attribute dictionary containing all of the group
192		instances contained within that group. Each [Group](#Group) instance also has a
	164	instances contained within that group. Each `Group` instance also has a
193	165	`path` attribute that contains a simulated unix directory path to
194	166	that group. To simplify the creation of nested groups, you can
195		use a unix-like path as an argument to [Dataset.createGroup](#Dataset.createGroup).
	167	use a unix-like path as an argument to `Dataset.createGroup`.
196	168
197	169	```python
198	170	>>> fcstgrp1 = rootgrp.createGroup("/forecasts/model1")

205	177	returned.
206	178
207	179	Here's an example that shows how to navigate all the groups in a
208		[Dataset](#Dataset). The function `walktree` is a Python generator that is used
209		to walk the directory tree. Note that printing the [Dataset](#Dataset) or [Group](#Group)
	180	`Dataset`. The function `walktree` is a Python generator that is used
	181	to walk the directory tree. Note that printing the `Dataset` or `Group`
210	182	object yields summary information about it's contents.
211	183
212	184	```python

253	225	before any variables can be created the dimensions they use must be
254	226	created first. A special case, not often used in practice, is that of a
255	227	scalar variable, which has no dimensions. A dimension is created using
256		the [Dataset.createDimension](#Dataset.createDimension) method of a [Dataset](#Dataset)
257		or [Group](#Group) instance. A Python string is used to set the name of the
	228	the `Dataset.createDimension` method of a `Dataset`
	229	or `Group` instance. A Python string is used to set the name of the
258	230	dimension, and an integer value is used to set the size. To create an
259	231	unlimited dimension (a dimension that can be appended to), the size
260	232	value is set to `None` or 0. In this example, there both the `time` and

270	242	```
271	243
272	244
273		All of the [Dimension](#Dimension) instances are stored in a python dictionary.
	245	All of the `Dimension` instances are stored in a python dictionary.
274	246
275	247	```python
276	248	>>> print(rootgrp.dimensions)
277	249	{'level': <class 'netCDF4._netCDF4.Dimension'> (unlimited): name = 'level', size = 0, 'time': <class 'netCDF4._netCDF4.Dimension'> (unlimited): name = 'time', size = 0, 'lat': <class 'netCDF4._netCDF4.Dimension'>: name = 'lat', size = 73, 'lon': <class 'netCDF4._netCDF4.Dimension'>: name = 'lon', size = 144}
278	250	```
279	251
280		Using the python `len` function with a [Dimension](#Dimension) instance returns
	252	Using the python `len` function with a `Dimension` instance returns
281	253	current size of that dimension.
282		[Dimension.isunlimited](#Dimension.isunlimited) method of a [Dimension](#Dimension) instance
	254	`Dimension.isunlimited` method of a `Dimension` instance
283	255	be used to determine if the dimensions is unlimited, or appendable.
284	256
285	257	```python

291	263	True
292	264	```
293	265
294		Printing the [Dimension](#Dimension) object
	266	Printing the `Dimension` object
295	267	provides useful summary info, including the name and length of the dimension,
296	268	and whether it is unlimited.
297	269

304	276	<class 'netCDF4._netCDF4.Dimension'>: name = 'lon', size = 144
305	277	```
306	278
307		[Dimension](#Dimension) names can be changed using the
308		[Datatset.renameDimension](#Datatset.renameDimension) method of a [Dataset](#Dataset) or
309		[Group](#Group) instance.
	279	`Dimension` names can be changed using the
	280	`Datatset.renameDimension` method of a `Dataset` or
	281	`Group` instance.
310	282
311	283	## Variables in a netCDF file
312	284

314	286	supplied by the [numpy module](http://numpy.scipy.org). However,
315	287	unlike numpy arrays, netCDF4 variables can be appended to along one or
316	288	more 'unlimited' dimensions. To create a netCDF variable, use the
317		[Dataset.createVariable](#Dataset.createVariable) method of a [Dataset](#Dataset) or
318		[Group](#Group) instance. The [Dataset.createVariable](#Dataset.createVariable) method
	289	`Dataset.createVariable` method of a `Dataset` or
	290	`Group` instance. The `Dataset.createVariable`j method
319	291	has two mandatory arguments, the variable name (a Python string), and
320	292	the variable datatype. The variable's dimensions are given by a tuple
321	293	containing the dimension names (defined previously with
322		[Dataset.createDimension](#Dataset.createDimension)). To create a scalar
	294	`Dataset.createDimension`). To create a scalar
323	295	variable, simply leave out the dimensions keyword. The variable
324	296	primitive datatypes correspond to the dtype attribute of a numpy array.
325	297	You can specify the datatype as a numpy dtype object, or anything that

337	309	can only be used if the file format is `NETCDF4`.
338	310
339	311	The dimensions themselves are usually also defined as variables, called
340		coordinate variables. The [Dataset.createVariable](#Dataset.createVariable)
341		method returns an instance of the [Variable](#Variable) class whose methods can be
	312	coordinate variables. The `Dataset.createVariable`
	313	method returns an instance of the `Variable` class whose methods can be
342	314	used later to access and set variable data and attributes.
343	315
344	316	```python

351	323	>>> temp.units = "K"
352	324	```
353	325
354		To get summary info on a [Variable](#Variable) instance in an interactive session,
	326	To get summary info on a `Variable` instance in an interactive session,
355	327	just print it.
356	328
357	329	```python

372	344
373	345	If the intermediate groups do not yet exist, they will be created.
374	346
375		You can also query a [Dataset](#Dataset) or [Group](#Group) instance directly to obtain [Group](#Group) or
376		[Variable](#Variable) instances using paths.
	347	You can also query a `Dataset` or `Group` instance directly to obtain `Group` or
	348	`Variable` instances using paths.
377	349
378	350	```python
379	351	>>> print(rootgrp["/forecasts/model1"]) # a Group instance

392	364	```
393	365
394	366
395		All of the variables in the [Dataset](#Dataset) or [Group](#Group) are stored in a
	367	All of the variables in the `Dataset` or `Group` are stored in a
396	368	Python dictionary, in the same way as the dimensions:
397	369
398	370	```python

421	393	filling on, default _FillValue of 9.969209968386869e+36 used}
422	394	```
423	395
424		[Variable](#Variable) names can be changed using the
425		[Dataset.renameVariable](#Dataset.renameVariable) method of a [Dataset](#Dataset)
	396	`Variable` names can be changed using the
	397	`Dataset.renameVariable` method of a `Dataset`
426	398	instance.
427	399
428	400	Variables can be sliced similar to numpy arrays, but there are some differences. See

433	405
434	406	There are two types of attributes in a netCDF file, global and variable.
435	407	Global attributes provide information about a group, or the entire
436		dataset, as a whole. [Variable](#Variable) attributes provide information about
	408	dataset, as a whole. `Variable` attributes provide information about
437	409	one of the variables in a group. Global attributes are set by assigning
438		values to [Dataset](#Dataset) or [Group](#Group) instance variables. [Variable](#Variable)
439		attributes are set by assigning values to [Variable](#Variable) instances
	410	values to `Dataset` or `Group` instance variables. `Variable`
	411	attributes are set by assigning values to `Variable` instances
440	412	variables. Attributes can be strings, numbers or sequences. Returning to
441	413	our example,
442	414

453	425	>>> times.calendar = "gregorian"
454	426	```
455	427
456		The [Dataset.ncattrs](#Dataset.ncattrs) method of a [Dataset](#Dataset), [Group](#Group) or
457		[Variable](#Variable) instance can be used to retrieve the names of all the netCDF
	428	The `Dataset.ncattrs` method of a `Dataset`, `Group` or
	429	`Variable` instance can be used to retrieve the names of all the netCDF
458	430	attributes. This method is provided as a convenience, since using the
459	431	built-in `dir` Python function will return a bunch of private methods
460	432	and attributes that cannot (or should not) be modified by the user.

467	439	Global attr source = netCDF4 python module tutorial
468	440	```
469	441
470		The `__dict__` attribute of a [Dataset](#Dataset), [Group](#Group) or [Variable](#Variable)
	442	The `__dict__` attribute of a `Dataset`, `Group` or `Variable`
471	443	instance provides all the netCDF attribute name/value pairs in a python
472	444	dictionary:
473	445

476	448	{'description': 'bogus example script', 'history': 'Created Mon Jul 8 14:19:41 2019', 'source': 'netCDF4 python module tutorial'}
477	449	```
478	450
479		Attributes can be deleted from a netCDF [Dataset](#Dataset), [Group](#Group) or
480		[Variable](#Variable) using the python `del` statement (i.e. `del grp.foo`
	451	Attributes can be deleted from a netCDF `Dataset`, `Group` or
	452	`Variable` using the python `del` statement (i.e. `del grp.foo`
481	453	removes the attribute `foo` the the group `grp`).
482	454
483	455	## Writing data to and retrieving data from a netCDF variable
484	456
485		Now that you have a netCDF [Variable](#Variable) instance, how do you put data
	457	Now that you have a netCDF `Variable` instance, how do you put data
486	458	into it? You can just treat it like an array and assign data to a slice.
487	459
488	460	```python
489		>>> import numpy
490		>>> lats = numpy.arange(-90,91,2.5)
491		>>> lons = numpy.arange(-180,180,2.5)
	461	>>> import numpy as np
	462	>>> lats = np.arange(-90,91,2.5)
	463	>>> lons = np.arange(-180,180,2.5)
492	464	>>> latitudes[:] = lats
493	465	>>> longitudes[:] = lons
494	466	>>> print("latitudes =\\n{}".format(latitudes[:]))

502	474	90. ]
503	475	```
504	476
505		Unlike NumPy's array objects, netCDF [Variable](#Variable)
	477	Unlike NumPy's array objects, netCDF `Variable`
506	478	objects with unlimited dimensions will grow along those dimensions if you
507	479	assign data outside the currently defined range of indices.
508	480

579	551
580	552	By default, netcdf4-python returns numpy masked arrays with values equal to the
581	553	`missing_value` or `_FillValue` variable attributes masked. The
582		[Dataset.set_auto_mask](#Dataset.set_auto_mask) [Dataset](#Dataset) and [Variable](#Variable) methods
	554	`Dataset.set_auto_mask` `Dataset` and `Variable` methods
583	555	can be used to disable this feature so that
584	556	numpy arrays are always returned, with the missing values included. Prior to
585	557	version 1.4.0 the default behavior was to only return masked arrays when the
586	558	requested slice contained missing values. This behavior can be recovered
587		using the [Dataset.set_always_mask](#Dataset.set_always_mask) method. If a masked array is
	559	using the `Dataset.set_always_mask` method. If a masked array is
588	560	written to a netCDF variable, the masked elements are filled with the
589	561	value specified by the `missing_value` attribute. If the variable has
590	562	no `missing_value`, the `_FillValue` is used instead.

629	601	## Reading data from a multi-file netCDF dataset
630	602
631	603	If you want to read data from a variable that spans multiple netCDF files,
632		you can use the [MFDataset](#MFDataset) class to read the data as if it were
	604	you can use the `MFDataset` class to read the data as if it were
633	605	contained in a single file. Instead of using a single filename to create
634		a [Dataset](#Dataset) instance, create a [MFDataset](#MFDataset) instance with either a list
	606	a `Dataset` instance, create a `MFDataset` instance with either a list
635	607	of filenames, or a string with a wildcard (which is then converted to
636	608	a sorted list of files using the python glob module).
637	609	Variables in the list of files that share the same unlimited

647	619	... with Dataset("mftest%s.nc" % nf, "w", format="NETCDF4_CLASSIC") as f:
648	620	... _ = f.createDimension("x",None)
649	621	... x = f.createVariable("x","i",("x",))
650		... x[0:10] = numpy.arange(nf10,10(nf+1))
	622	... x[0:10] = np.arange(nf10,10(nf+1))
651	623	```
652	624
653		Now read all the files back in at once with [MFDataset](#MFDataset)
	625	Now read all the files back in at once with `MFDataset`
654	626
655	627	```python
656	628	>>> from netCDF4 import MFDataset

663	635	96 97 98 99]
664	636	```
665	637
666		Note that [MFDataset](#MFDataset) can only be used to read, not write, multi-file
	638	Note that `MFDataset` can only be used to read, not write, multi-file
667	639	datasets.
668	640
669	641	## Efficient compression of netCDF variables
670	642
671		Data stored in netCDF 4 [Variable](#Variable) objects can be compressed and
	643	Data stored in netCDF 4 `Variable` objects can be compressed and
672	644	decompressed on the fly. The parameters for the compression are
673	645	determined by the `zlib`, `complevel` and `shuffle` keyword arguments
674		to the [Dataset.createVariable](#Dataset.createVariable) method. To turn on
	646	to the `Dataset.createVariable` method. To turn on
675	647	compression, set `zlib=True`. The `complevel` keyword regulates the
676	648	speed and efficiency of the compression (1 being fastest, but lowest
677	649	compression ratio, 9 being slowest but best compression ratio). The

680	652	compression by reordering the bytes. The shuffle filter can
681	653	significantly improve compression ratios, and is on by default. Setting
682	654	`fletcher32` keyword argument to
683		[Dataset.createVariable](#Dataset.createVariable) to `True` (it's `False` by
	655	`Dataset.createVariable` to `True` (it's `False` by
684	656	default) enables the Fletcher32 checksum algorithm for error detection.
685	657	It's also possible to set the HDF5 chunking parameters and endian-ness
686	658	of the binary data stored in the HDF5 file with the `chunksizes`
687	659	and `endian` keyword arguments to
688		[Dataset.createVariable](#Dataset.createVariable). These keyword arguments only
	660	`Dataset.createVariable`. These keyword arguments only
689	661	are relevant for `NETCDF4` and `NETCDF4_CLASSIC` files (where the
690	662	underlying file format is HDF5) and are silently ignored if the file
691	663	format is `NETCDF3_CLASSIC`, `NETCDF3_64BIT_OFFSET` or `NETCDF3_64BIT_DATA`.

694	666	example, it is temperature data that was measured with a precision of
695	667	0.1 degrees), you can dramatically improve zlib compression by
696	668	quantizing (or truncating) the data using the `least_significant_digit`
697		keyword argument to [Dataset.createVariable](#Dataset.createVariable). The least
	669	keyword argument to `Dataset.createVariable`. The least
698	670	significant digit is the power of ten of the smallest decimal place in
699	671	the data that is a reliable value. For example if the data has a
700	672	precision of 0.1, then setting `least_significant_digit=1` will cause

706	678
707	679	In our example, try replacing the line
708	680
709		```python
710		>>> temp = rootgrp.createVariable("temp","f4",("time","level","lat","lon",))
	681	```python
	682	>>> temp = rootgrp.createVariable("temp","f4",("time","level","lat","lon",))
	683	```
711	684
712	685	with
713	686

735	708	information for a point by reading one variable, instead of reading
736	709	different parameters from different variables. Compound data types
737	710	are created from the corresponding numpy data type using the
738		[Dataset.createCompoundType](#Dataset.createCompoundType) method of a [Dataset](#Dataset) or [Group](#Group) instance.
	711	`Dataset.createCompoundType` method of a `Dataset` or `Group` instance.
739	712	Since there is no native complex data type in netcdf, compound types are handy
740	713	for storing numpy complex arrays. Here's an example:
741	714

743	716	>>> f = Dataset("complex.nc","w")
744	717	>>> size = 3 # length of 1-d complex array
745	718	>>> # create sample complex data.
746		>>> datac = numpy.exp(1j*(1.+numpy.linspace(0, numpy.pi, size)))
	719	>>> datac = np.exp(1j*(1.+np.linspace(0, np.pi, size)))
747	720	>>> # create complex128 compound data type.
748		>>> complex128 = numpy.dtype([("real",numpy.float_64),("imag",numpy.float_64)])
	721	>>> complex128 = np.dtype([("real",np.float64),("imag",np.float64)])
749	722	>>> complex128_t = f.createCompoundType(complex128,"complex128")
750	723	>>> # create a variable with this data type, write some data to it.
751	724	>>> x_dim = f.createDimension("x_dim",None)
752	725	>>> v = f.createVariable("cmplx_var",complex128_t,"x_dim")
753		>>> data = numpy.empty(size,complex128) # numpy structured array
	726	>>> data = np.empty(size,complex128) # numpy structured array
754	727	>>> data["real"] = datac.real; data["imag"] = datac.imag
755	728	>>> v[:] = data # write numpy structured array to netcdf compound var
756	729	>>> # close and reopen the file, check the contents.

758	731	>>> v = f.variables["cmplx_var"]
759	732	>>> datain = v[:] # read in all the data into a numpy structured array
760	733	>>> # create an empty numpy complex array
761		>>> datac2 = numpy.empty(datain.shape,numpy.complex128)
	734	>>> datac2 = np.empty(datain.shape,np.complex128)
762	735	>>> # .. fill it with contents of structured array.
763	736	>>> datac2.real = datain["real"]; datac2.imag = datain["imag"]
764	737	>>> print('{}: {}'.format(datac.dtype, datac)) # original data

772	745	ones first. All possible numpy structured arrays cannot be
773	746	represented as Compound variables - an error message will be
774	747	raise if you try to create one that is not supported.
775		All of the compound types defined for a [Dataset](#Dataset) or [Group](#Group) are stored
	748	All of the compound types defined for a `Dataset` or `Group` are stored
776	749	in a Python dictionary, just like variables and dimensions. As always, printing
777	750	objects gives useful summary information in an interactive session:
778	751

799	772
800	773	NetCDF 4 has support for variable-length or "ragged" arrays. These are arrays
801	774	of variable length sequences having the same type. To create a variable-length
802		data type, use the [Dataset.createVLType](#Dataset.createVLType) method
803		method of a [Dataset](#Dataset) or [Group](#Group) instance.
	775	data type, use the `Dataset.createVLType` method
	776	method of a `Dataset` or `Group` instance.
804	777
805	778	```python
806	779	>>> f = Dataset("tst_vlen.nc","w")
807		>>> vlen_t = f.createVLType(numpy.int32, "phony_vlen")
	780	>>> vlen_t = f.createVLType(np.int32, "phony_vlen")
808	781	```
809	782
810	783	The numpy datatype of the variable-length sequences and the name of the

830	803	```python
831	804	>>> import random
832	805	>>> random.seed(54321)
833		>>> data = numpy.empty(len(y)*len(x),object)
	806	>>> data = np.empty(len(y)*len(x),object)
834	807	>>> for n in range(len(y)*len(x)):
835		... data[n] = numpy.arange(random.randint(1,10),dtype="int32")+1
836		>>> data = numpy.reshape(data,(len(y),len(x)))
	808	... data[n] = np.arange(random.randint(1,10),dtype="int32")+1
	809	>>> data = np.reshape(data,(len(y),len(x)))
837	810	>>> vlvar[:] = data
838	811	>>> print("vlen variable =\\n{}".format(vlvar[:]))
839	812	vlen variable =

866	839	variables, For vlen strings, you don't need to create a vlen data type.
867	840	Instead, simply use the python `str` builtin (or a numpy string datatype
868	841	with fixed length greater than 1) when calling the
869		[Dataset.createVariable](#Dataset.createVariable) method.
	842	`Dataset.createVariable` method.
870	843
871	844	```python
872	845	>>> z = f.createDimension("z",10)

879	852
880	853	```python
881	854	>>> chars = "1234567890aabcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"
882		>>> data = numpy.empty(10,"O")
	855	>>> data = np.empty(10,"O")
883	856	>>> for n in range(10):
884	857	... stringlen = random.randint(2,12)
885	858	... data[n] = "".join([random.choice(chars) for i in range(stringlen)])

915	888	Here's an example of using an Enum type to hold cloud type data.
916	889	The base integer data type and a python dictionary describing the allowed
917	890	values and their names are used to define an Enum data type using
918		[Dataset.createEnumType](#Dataset.createEnumType).
	891	`Dataset.createEnumType`.
919	892
920	893	```python
921	894	>>> nc = Dataset('clouds.nc','w')

925	898	... 'Nimbostratus': 6, 'Cumulus': 4, 'Altostratus': 5,
926	899	... 'Cumulonimbus': 1, 'Stratocumulus': 3}
927	900	>>> # create the Enum type called 'cloud_t'.
928		>>> cloud_type = nc.createEnumType(numpy.uint8,'cloud_t',enum_dict)
	901	>>> cloud_type = nc.createEnumType(np.uint8,'cloud_t',enum_dict)
929	902	>>> print(cloud_type)
930	903	<class 'netCDF4._netCDF4.EnumType'>: name = 'cloud_t', numpy dtype = uint8, fields/values ={'Altocumulus': 7, 'Missing': 255, 'Stratus': 2, 'Clear': 0, 'Nimbostratus': 6, 'Cumulus': 4, 'Altostratus': 5, 'Cumulonimbus': 1, 'Stratocumulus': 3}
931	904	```

998	971
999	972	```python
1000	973	>>> d = nc.createDimension('dim',4)
1001		>>> v = nc.createVariable('var', np.int, 'dim')
	974	>>> v = nc.createVariable('var', np.int64, 'dim')
1002	975	>>> v[rank] = rank
1003	976	>>> nc.close()
1004	977

1019	992	depend on or be affected by other processes. Collective IO is a way of doing
1020	993	IO defined in the MPI-IO standard; unlike independent IO, all processes must
1021	994	participate in doing IO. To toggle back and forth between
1022		the two types of IO, use the [Variable.set_collective](#Variable.set_collective)
1023		[Variable](#Variable) method. All metadata
	995	the two types of IO, use the `Variable.set_collective`
	996	`Variable` method. All metadata
1024	997	operations (such as creation of groups, types, variables, dimensions, or attributes)
1025	998	are collective. There are a couple of important limitations of parallel IO:
1026	999

1032	1005	If the write is done in independent mode, the operation will fail with a
1033	1006	a generic "HDF Error".
1034	1007	- You can write compressed data in parallel only with netcdf-c >= 4.7.4
1035		and hdf5 >= 1.10.3 (although you can read in parallel with earlier versions).
	1008	and hdf5 >= 1.10.3 (although you can read in parallel with earlier versions). To write
	1009	compressed data in parallel, the variable must be in 'collective IO mode'. This is done
	1010	automatically on variable creation if compression is turned on, but if you are appending
	1011	to a variable in an existing file, you must use `Variable.set_collective(True)` before attempting
	1012	to write to it.
1036	1013	- You cannot use variable-length (VLEN) data types.
1037	1014
1038	1015	## Dealing with strings

1048	1025	To perform the conversion to and from character arrays to fixed-width numpy string arrays, the
1049	1026	following convention is followed by the python interface.
1050	1027	If the `_Encoding` special attribute is set for a character array
1051		(dtype `S1`) variable, the [chartostring](#chartostring) utility function is used to convert the array of
	1028	(dtype `S1`) variable, the `chartostring` utility function is used to convert the array of
1052	1029	characters to an array of strings with one less dimension (the last dimension is
1053	1030	interpreted as the length of each string) when reading the data. The character
1054	1031	set (usually ascii) is specified by the `_Encoding` attribute. If `_Encoding`
1055	1032	is 'none' or 'bytes', then the character array is converted to a numpy
1056	1033	fixed-width byte string array (dtype `S#`), otherwise a numpy unicode (dtype
1057	1034	`U#`) array is created. When writing the data,
1058		[stringtochar](#stringtochar) is used to convert the numpy string array to an array of
	1035	`stringtochar` is used to convert the numpy string array to an array of
1059	1036	characters with one more dimension. For example,
1060	1037
1061	1038	```python

1064	1041	>>> _ = nc.createDimension('nchars',3)
1065	1042	>>> _ = nc.createDimension('nstrings',None)
1066	1043	>>> v = nc.createVariable('strings','S1',('nstrings','nchars'))
1067		>>> datain = numpy.array(['foo','bar'],dtype='S3')
	1044	>>> datain = np.array(['foo','bar'],dtype='S3')
1068	1045	>>> v[:] = stringtochar(datain) # manual conversion to char array
1069	1046	>>> print(v[:]) # data returned as char array
1070	1047	[[b'f' b'o' b'o']

1078	1055
1079	1056	Even if the `_Encoding` attribute is set, the automatic conversion of char
1080	1057	arrays to/from string arrays can be disabled with
1081		[Variable.set_auto_chartostring](#Variable.set_auto_chartostring).
	1058	`Variable.set_auto_chartostring`.
1082	1059
1083	1060	A similar situation is often encountered with numpy structured arrays with subdtypes
1084	1061	containing fixed-wdith byte strings (dtype=`S#`). Since there is no native fixed-length string

1094	1071
1095	1072	```python
1096	1073	>>> nc = Dataset('compoundstring_example.nc','w')
1097		>>> dtype = numpy.dtype([('observation', 'f4'),
	1074	>>> dtype = np.dtype([('observation', 'f4'),
1098	1075	... ('station_name','S10')])
1099	1076	>>> station_data_t = nc.createCompoundType(dtype,'station_data')
1100	1077	>>> _ = nc.createDimension('station',None)
1101	1078	>>> statdat = nc.createVariable('station_obs', station_data_t, ('station',))
1102		>>> data = numpy.empty(2,dtype)
	1079	>>> data = np.empty(2,dtype)
1103	1080	>>> data['observation'][:] = (123.,3.14)
1104	1081	>>> data['station_name'][:] = ('Boulder','New York')
1105	1082	>>> print(statdat.dtype) # strings actually stored as character arrays

1143	1120	>>> # and persist the file to disk when it is closed.
1144	1121	>>> nc = Dataset('diskless_example.nc','w',diskless=True,persist=True)
1145	1122	>>> d = nc.createDimension('x',None)
1146		>>> v = nc.createVariable('v',numpy.int32,'x')
1147		>>> v[0:5] = numpy.arange(5)
	1123	>>> v = nc.createVariable('v',np.int32,'x')
	1124	>>> v[0:5] = np.arange(5)
1148	1125	>>> print(nc)
1149	1126	<class 'netCDF4._netCDF4.Dataset'>
1150	1127	root group (NETCDF4 data model, file format HDF5):

1177	1154	>>> # (ignored for NETCDF4/HDF5 files).
1178	1155	>>> nc = Dataset('inmemory.nc', mode='w',memory=1028)
1179	1156	>>> d = nc.createDimension('x',None)
1180		>>> v = nc.createVariable('v',numpy.int32,'x')
1181		>>> v[0:5] = numpy.arange(5)
	1157	>>> v = nc.createVariable('v',np.int32,'x')
	1158	>>> v[0:5] = np.arange(5)
1182	1159	>>> nc_buf = nc.close() # close returns memoryview
1183	1160	>>> print(type(nc_buf))
1184	1161	<class 'memoryview'>

1226	1203	# Python 3.7+ guarantees order; older versions need OrderedDict
1227	1204	from collections import OrderedDict
1228	1205
1229		__version__ = "1.5.6"
	1206	__version__ = "1.5.7"
1230	1207
1231	1208	# Initialize numpy
1232	1209	import posixpath

1280	1257
1281	1258	return current netCDF chunk cache information in a tuple (size,nelems,preemption).
1282	1259	See netcdf C library documentation for `nc_get_chunk_cache` for
1283		details. Values can be reset with [set_chunk_cache](#set_chunk_cache)."""
	1260	details. Values can be reset with `set_chunk_cache`."""
1284	1261	cdef int ierr
1285	1262	cdef size_t sizep, nelemsp
1286	1263	cdef float preemptionp

1679	1656	_ensure_nc_success(ierr, err_cls=AttributeError)
1680	1657
1681	1658	cdef _get_types(group):
1682		# Private function to create [CompoundType](#CompoundType),
1683		# [VLType](#VLType) or [EnumType](#EnumType) instances for all the
1684		# compound, VLEN or Enum types in a [Group](#Group) or [Dataset](#Dataset).
	1659	# Private function to create `CompoundType`,
	1660	# `VLType` or `EnumType` instances for all the
	1661	# compound, VLEN or Enum types in a `Group` or `Dataset`.
1685	1662	cdef int ierr, ntypes, classp, n, _grpid
1686	1663	cdef nc_type xtype
1687	1664	cdef nc_type *typeids

1750	1727	return cmptypes, vltypes, enumtypes
1751	1728
1752	1729	cdef _get_dims(group):
1753		# Private function to create [Dimension](#Dimension) instances for all the
1754		# dimensions in a [Group](#Group) or Dataset
	1730	# Private function to create `Dimension` instances for all the
	1731	# dimensions in a `Group` or Dataset
1755	1732	cdef int ierr, numdims, n, _grpid
1756	1733	cdef int *dimids
1757	1734	cdef char namstring[NC_MAX_NAME+1]

1784	1761	return dimensions
1785	1762
1786	1763	cdef _get_grps(group):
1787		# Private function to create [Group](#Group) instances for all the
1788		# groups in a [Group](#Group) or Dataset
	1764	# Private function to create `Group` instances for all the
	1765	# groups in a `Group` or Dataset
1789	1766	cdef int ierr, numgrps, n, _grpid
1790	1767	cdef int *grpids
1791	1768	cdef char namstring[NC_MAX_NAME+1]

1794	1771	with nogil:
1795	1772	ierr = nc_inq_grps(_grpid, &numgrps, NULL)
1796	1773	_ensure_nc_success(ierr)
1797		# create dictionary containing [Group](#Group) instances for groups in this group
	1774	# create dictionary containing `Group` instances for groups in this group
1798	1775	if sys.version_info[0:2] < (3, 7):
1799	1776	groups = OrderedDict()
1800	1777	else:

1814	1791	return groups
1815	1792
1816	1793	cdef _get_vars(group):
1817		# Private function to create [Variable](#Variable) instances for all the
1818		# variables in a [Group](#Group) or Dataset
	1794	# Private function to create `Variable` instances for all the
	1795	# variables in a `Group` or Dataset
1819	1796	cdef int ierr, numvars, n, nn, numdims, varid, classp, iendian, _grpid
1820	1797	cdef int *varids
1821	1798	cdef int *dimids

1960	1937
1961	1938	cdef class Dataset:
1962	1939	"""
1963		A netCDF [Dataset](#Dataset) is a collection of dimensions, groups, variables and
	1940	A netCDF `Dataset` is a collection of dimensions, groups, variables and
1964	1941	attributes. Together they describe the meaning of data and relations among
1965		data fields stored in a netCDF file. See [Dataset.__init__](#Dataset.__init__) for more
	1942	data fields stored in a netCDF file. See `Dataset.__init__` for more
1966	1943	details.
1967	1944
1968	1945	A list of attribute names corresponding to global netCDF attributes
1969		defined for the [Dataset](#Dataset) can be obtained with the
1970		[Dataset.ncattrs](#Dataset.ncattrs) method.
	1946	defined for the `Dataset` can be obtained with the
	1947	`Dataset.ncattrs` method.
1971	1948	These attributes can be created by assigning to an attribute of the
1972		[Dataset](#Dataset) instance. A dictionary containing all the netCDF attribute
	1949	`Dataset` instance. A dictionary containing all the netCDF attribute
1973	1950	name/value pairs is provided by the `__dict__` attribute of a
1974		[Dataset](#Dataset) instance.
	1951	`Dataset` instance.
1975	1952
1976	1953	The following class variables are read-only and should not be
1977	1954	modified by the user.
1978	1955
1979	1956	`dimensions`: The `dimensions` dictionary maps the names of
1980		dimensions defined for the [Group](#Group) or [Dataset](#Dataset) to instances of the
1981		[Dimension](#Dimension) class.
	1957	dimensions defined for the `Group` or `Dataset` to instances of the
	1958	`Dimension` class.
1982	1959
1983	1960	`variables`: The `variables` dictionary maps the names of variables
1984		defined for this [Dataset](#Dataset) or [Group](#Group) to instances of the
1985		[Variable](#Variable) class.
	1961	defined for this `Dataset` or `Group` to instances of the
	1962	`Variable` class.
1986	1963
1987	1964	`groups`: The groups dictionary maps the names of groups created for
1988		this [Dataset](#Dataset) or [Group](#Group) to instances of the [Group](#Group) class (the
1989		[Dataset](#Dataset) class is simply a special case of the [Group](#Group) class which
	1965	this `Dataset` or `Group` to instances of the `Group` class (the
	1966	`Dataset` class is simply a special case of the `Group` class which
1990	1967	describes the root group in the netCDF4 file).
1991	1968
1992	1969	`cmptypes`: The `cmptypes` dictionary maps the names of
1993		compound types defined for the [Group](#Group) or [Dataset](#Dataset) to instances of the
1994		[CompoundType](#CompoundType) class.
	1970	compound types defined for the `Group` or `Dataset` to instances of the
	1971	`CompoundType` class.
1995	1972
1996	1973	`vltypes`: The `vltypes` dictionary maps the names of
1997		variable-length types defined for the [Group](#Group) or [Dataset](#Dataset) to instances
1998		of the [VLType](#VLType) class.
	1974	variable-length types defined for the `Group` or `Dataset` to instances
	1975	of the `VLType` class.
1999	1976
2000	1977	`enumtypes`: The `enumtypes` dictionary maps the names of
2001		Enum types defined for the [Group](#Group) or [Dataset](#Dataset) to instances
2002		of the [EnumType](#EnumType) class.
	1978	Enum types defined for the `Group` or `Dataset` to instances
	1979	of the `EnumType` class.
2003	1980
2004	1981	`data_model`: `data_model` describes the netCDF
2005	1982	data model version, one of `NETCDF3_CLASSIC`, `NETCDF4`,

2014	1991	`UNDEFINED`.
2015	1992
2016	1993	`parent`: `parent` is a reference to the parent
2017		[Group](#Group) instance. `None` for the root group or [Dataset](#Dataset)
	1994	`Group` instance. `None` for the root group or `Dataset`
2018	1995	instance.
2019	1996
2020		`path`: `path` shows the location of the [Group](#Group) in
2021		the [Dataset](#Dataset) in a unix directory format (the names of groups in the
2022		hierarchy separated by backslashes). A [Dataset](#Dataset) instance is the root
	1997	`path`: `path` shows the location of the `Group` in
	1998	the `Dataset` in a unix directory format (the names of groups in the
	1999	hierarchy separated by backslashes). A `Dataset` instance is the root
2023	2000	group, so the path is simply `'/'`.
2024	2001
2025	2002	`keepweakref`: If `True`, child Dimension and Variables objects only keep weak

2045	2022	persist=False, keepweakref=False, memory=None, encoding=None,
2046	2023	parallel=False, comm=None, info=None, format='NETCDF4')`**
2047	2024
2048		[Dataset](#Dataset) constructor.
	2025	`Dataset` constructor.
2049	2026
2050	2027	`filename`: Name of netCDF file to hold dataset. Can also
2051	2028	be a python 3 pathlib instance or the URL of an OpenDAP dataset. When memory is

2509	2486	"""
2510	2487	`sync(self)`
2511	2488
2512		Writes all buffered data in the [Dataset](#Dataset) to the disk file."""
	2489	Writes all buffered data in the `Dataset` to the disk file."""
2513	2490	_ensure_nc_success(nc_sync(self._grpid))
2514	2491
2515	2492	def _redef(self):

2524	2501	"""
2525	2502	`set_fill_on(self)`
2526	2503
2527		Sets the fill mode for a [Dataset](#Dataset) open for writing to `on`.
	2504	Sets the fill mode for a `Dataset` open for writing to `on`.
2528	2505
2529	2506	This causes data to be pre-filled with fill values. The fill values can be
2530	2507	controlled by the variable's `_Fill_Value` attribute, but is usually

2540	2517	"""
2541	2518	`set_fill_off(self)`
2542	2519
2543		Sets the fill mode for a [Dataset](#Dataset) open for writing to `off`.
	2520	Sets the fill mode for a `Dataset` open for writing to `off`.
2544	2521
2545	2522	This will prevent the data from being pre-filled with fill values, which
2546	2523	may result in some performance improvements. However, you must then make

2556	2533
2557	2534	`size` must be a positive integer or `None`, which stands for
2558	2535	"unlimited" (default is `None`). Specifying a size of 0 also
2559		results in an unlimited dimension. The return value is the [Dimension](#Dimension)
	2536	results in an unlimited dimension. The return value is the `Dimension`
2560	2537	class instance describing the new dimension. To determine the current
2561		maximum size of the dimension, use the `len` function on the [Dimension](#Dimension)
	2538	maximum size of the dimension, use the `len` function on the `Dimension`
2562	2539	instance. To determine if a dimension is 'unlimited', use the
2563		[Dimension.isunlimited](#Dimension.isunlimited) method of the [Dimension](#Dimension) instance."""
	2540	`Dimension.isunlimited` method of the `Dimension` instance."""
2564	2541	self.dimensions[dimname] = Dimension(self, dimname, size=size)
2565	2542	return self.dimensions[dimname]
2566	2543

2568	2545	"""
2569	2546	`renameDimension(self, oldname, newname)`
2570	2547
2571		rename a [Dimension](#Dimension) named `oldname` to `newname`."""
	2548	rename a `Dimension` named `oldname` to `newname`."""
2572	2549	cdef char *namstring
2573	2550	bytestr = _strencode(newname)
2574	2551	namstring = bytestr

2600	2577	are homogeneous numeric data types), then the 'inner' compound types must be
2601	2578	created first.
2602	2579
2603		The return value is the [CompoundType](#CompoundType) class instance describing the new
	2580	The return value is the `CompoundType` class instance describing the new
2604	2581	datatype."""
2605	2582	self.cmptypes[datatype_name] = CompoundType(self, datatype,\
2606	2583	datatype_name)

2613	2590	Creates a new VLEN data type named `datatype_name` from a numpy
2614	2591	dtype object `datatype`.
2615	2592
2616		The return value is the [VLType](#VLType) class instance describing the new
	2593	The return value is the `VLType` class instance describing the new
2617	2594	datatype."""
2618	2595	self.vltypes[datatype_name] = VLType(self, datatype, datatype_name)
2619	2596	return self.vltypes[datatype_name]

2626	2603	integer dtype object `datatype`, and a python dictionary
2627	2604	defining the enum fields and values.
2628	2605
2629		The return value is the [EnumType](#EnumType) class instance describing the new
	2606	The return value is the `EnumType` class instance describing the new
2630	2607	datatype."""
2631	2608	self.enumtypes[datatype_name] = EnumType(self, datatype, datatype_name,
2632	2609	enum_dict)

2657	2634	(NC_BYTE), 'u1' (NC_UBYTE), 'i2' or 'h' or 's' (NC_SHORT), 'u2'
2658	2635	(NC_USHORT), 'i4' or 'i' or 'l' (NC_INT), 'u4' (NC_UINT), 'i8' (NC_INT64),
2659	2636	'u8' (NC_UINT64), 'f4' or 'f' (NC_FLOAT), 'f8' or 'd' (NC_DOUBLE)`.
2660		`datatype` can also be a [CompoundType](#CompoundType) instance
2661		(for a structured, or compound array), a [VLType](#VLType) instance
	2637	`datatype` can also be a `CompoundType` instance
	2638	(for a structured, or compound array), a `VLType` instance
2662	2639	(for a variable-length array), or the python `str` builtin
2663	2640	(for a variable-length string array). Numpy string and unicode datatypes with
2664	2641	length greater than one are aliases for `str`.

2667	2644	the corresponding data type.
2668	2645
2669	2646	`dimensions` must be a tuple containing dimension names (strings) that
2670		have been defined previously using [Dataset.createDimension](#Dataset.createDimension). The default value
	2647	have been defined previously using `Dataset.createDimension`. The default value
2671	2648	is an empty tuple, which means the variable is a scalar.
2672	2649
2673	2650	If the optional keyword `zlib` is `True`, the data will be compressed in

2709	2686
2710	2687	The optional keyword `fill_value` can be used to override the default
2711	2688	netCDF `_FillValue` (the value that the variable gets filled with before
2712		any data is written to it, defaults given in [default_fillvals](#default_fillvals)).
	2689	any data is written to it, defaults given in the dict `netCDF4.default_fillvals`).
2713	2690	If fill_value is set to `False`, then the variable is not pre-filled.
2714	2691
2715	2692	If the optional keyword parameter `least_significant_digit` is

2734	2711	method to change it the next time the Dataset is opened.
2735	2712	Warning - messing with this parameter can seriously degrade performance.
2736	2713
2737		The return value is the [Variable](#Variable) class instance describing the new
	2714	The return value is the `Variable` class instance describing the new
2738	2715	variable.
2739	2716
2740	2717	A list of names corresponding to netCDF variable attributes can be
2741		obtained with the [Variable](#Variable) method [Variable.ncattrs](#Variable.ncattrs). A dictionary
	2718	obtained with the `Variable` method `Variable.ncattrs`. A dictionary
2742	2719	containing all the netCDF attribute name/value pairs is provided by
2743		the `__dict__` attribute of a [Variable](#Variable) instance.
2744
2745		[Variable](#Variable) instances behave much like array objects. Data can be
	2720	the `__dict__` attribute of a `Variable` instance.
	2721
	2722	`Variable` instances behave much like array objects. Data can be
2746	2723	assigned to or retrieved from a variable with indexing and slicing
2747		operations on the [Variable](#Variable) instance. A [Variable](#Variable) instance has six
	2724	operations on the `Variable` instance. A `Variable` instance has six
2748	2725	Dataset standard attributes: `dimensions, dtype, shape, ndim, name` and
2749	2726	`least_significant_digit`. Application programs should never modify
2750	2727	these attributes. The `dimensions` attribute is a tuple containing the

2755	2732	string containing the name of the Variable instance.
2756	2733	The `least_significant_digit`
2757	2734	attributes describes the power of ten of the smallest decimal place in
2758		the data the contains a reliable value. assigned to the [Variable](#Variable)
	2735	the data the contains a reliable value. assigned to the `Variable`
2759	2736	instance. If `None`, the data is not truncated. The `ndim` attribute
2760	2737	is the number of variable dimensions."""
2761	2738	# if varname specified as a path, split out group names.

2778	2755	"""
2779	2756	`renameVariable(self, oldname, newname)`
2780	2757
2781		rename a [Variable](#Variable) named `oldname` to `newname`"""
	2758	rename a `Variable` named `oldname` to `newname`"""
2782	2759	cdef char *namstring
2783	2760	try:
2784	2761	var = self.variables[oldname]

2800	2777	"""
2801	2778	`createGroup(self, groupname)`
2802	2779
2803		Creates a new [Group](#Group) with the given `groupname`.
	2780	Creates a new `Group` with the given `groupname`.
2804	2781
2805	2782	If `groupname` is specified as a path, using forward slashes as in unix to
2806	2783	separate components, then intermediate groups will be created as necessary

2810	2787	If the specified path describes a group that already exists, no error is
2811	2788	raised.
2812	2789
2813		The return value is a [Group](#Group) class instance."""
	2790	The return value is a `Group` class instance."""
2814	2791	# if group specified as a path, split out group names
2815	2792	groupname = posixpath.normpath(groupname)
2816	2793	nestedgroups = groupname.split('/')

2830	2807	"""
2831	2808	`ncattrs(self)`
2832	2809
2833		return netCDF global attribute names for this [Dataset](#Dataset) or [Group](#Group) in a list."""
	2810	return netCDF global attribute names for this `Dataset` or `Group` in a list."""
2834	2811	return _get_att_names(self._grpid, NC_GLOBAL)
2835	2812
2836	2813	def setncattr(self,name,value):

2947	2924	"""
2948	2925	`renameAttribute(self, oldname, newname)`
2949	2926
2950		rename a [Dataset](#Dataset) or [Group](#Group) attribute named `oldname` to `newname`."""
	2927	rename a `Dataset` or `Group` attribute named `oldname` to `newname`."""
2951	2928	cdef char *oldnamec
2952	2929	cdef char *newnamec
2953	2930	bytestr = _strencode(oldname)

2960	2937	"""
2961	2938	`renameGroup(self, oldname, newname)`
2962	2939
2963		rename a [Group](#Group) named `oldname` to `newname` (requires netcdf >= 4.3.1)."""
	2940	rename a `Group` named `oldname` to `newname` (requires netcdf >= 4.3.1)."""
2964	2941	cdef char *newnamec
2965	2942	IF HAS_RENAME_GRP:
2966	2943	bytestr = _strencode(newname)

2984	2961	"""
2985	2962	`set_auto_chartostring(self, True_or_False)`
2986	2963
2987		Call [Variable.set_auto_chartostring](#Variable.set_auto_chartostring) for all variables contained in this [Dataset](#Dataset) or
2988		[Group](#Group), as well as for all variables in all its subgroups.
	2964	Call `Variable.set_auto_chartostring` for all variables contained in this `Dataset` or
	2965	`Group`, as well as for all variables in all its subgroups.
2989	2966
2990	2967	`True_or_False`: Boolean determining if automatic conversion of
2991	2968	all character arrays <--> string arrays should be performed for

3012	2989	"""
3013	2990	`set_auto_maskandscale(self, True_or_False)`
3014	2991
3015		Call [Variable.set_auto_maskandscale](#Variable.set_auto_maskandscale) for all variables contained in this [Dataset](#Dataset) or
3016		[Group](#Group), as well as for all variables in all its subgroups.
	2992	Call `Variable.set_auto_maskandscale` for all variables contained in this `Dataset` or
	2993	`Group`, as well as for all variables in all its subgroups.
3017	2994
3018	2995	`True_or_False`: Boolean determining if automatic conversion to masked arrays
3019	2996	and variable scaling shall be applied for all variables.

3039	3016	"""
3040	3017	`set_auto_mask(self, True_or_False)`
3041	3018
3042		Call [Variable.set_auto_mask](#Variable.set_auto_mask) for all variables contained in this [Dataset](#Dataset) or
3043		[Group](#Group), as well as for all variables in all its subgroups.
	3019	Call `Variable.set_auto_mask` for all variables contained in this `Dataset` or
	3020	`Group`, as well as for all variables in all its subgroups.
3044	3021
3045	3022	`True_or_False`: Boolean determining if automatic conversion to masked arrays
3046	3023	shall be applied for all variables.

3065	3042	"""
3066	3043	`set_auto_scale(self, True_or_False)`
3067	3044
3068		Call [Variable.set_auto_scale](#Variable.set_auto_scale) for all variables contained in this [Dataset](#Dataset) or
3069		[Group](#Group), as well as for all variables in all its subgroups.
	3045	Call `Variable.set_auto_scale` for all variables contained in this `Dataset` or
	3046	`Group`, as well as for all variables in all its subgroups.
3070	3047
3071	3048	`True_or_False`: Boolean determining if automatic variable scaling
3072	3049	shall be applied for all variables.

3091	3068	"""
3092	3069	`set_always_mask(self, True_or_False)`
3093	3070
3094		Call [Variable.set_always_mask](#Variable.set_always_mask) for all variables contained in
3095		this [Dataset](#Dataset) or [Group](#Group), as well as for all
	3071	Call `Variable.set_always_mask` for all variables contained in
	3072	this `Dataset` or `Group`, as well as for all
3096	3073	variables in all its subgroups.
3097	3074
3098	3075	`True_or_False`: Boolean determining if automatic conversion of

3122	3099	"""
3123	3100	`set_ncstring_attrs(self, True_or_False)`
3124	3101
3125		Call [Variable.set_ncstring_attrs](#Variable.set_ncstring_attrs) for all variables contained in
3126		this [Dataset](#Dataset) or [Group](#Group), as well as for all its
	3102	Call `Variable.set_ncstring_attrs` for all variables contained in
	3103	this `Dataset` or `Group`, as well as for all its
3127	3104	subgroups and their variables.
3128	3105
3129	3106	`True_or_False`: Boolean determining if all string attributes are

3224	3201	"""
3225	3202	`fromcdl(cdlfilename, ncfilename=None, mode='a',format='NETCDF4')`
3226	3203
3227		call `ncgen` via subprocess to create Dataset from [CDL](https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#cdl_guide)
3228		text representation. Requires `ncgen` to be installed and in `$PATH`.
	3204	call [ncgen][ncgen] via subprocess to create Dataset from [CDL][cdl]
	3205	text representation. Requires [ncgen][ncgen] to be installed and in `$PATH`.
3229	3206
3230	3207	`cdlfilename`: CDL file.
3231	3208

3239	3216	`'NETCDF3_64BIT_DATA'`. Default `'NETCDF4'`.
3240	3217
3241	3218	Dataset instance for `ncfilename` is returned.
	3219
	3220	[ncgen]: https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#ncgen_guide
	3221	[cdl]: https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#cdl_guide
3242	3222	"""
3243	3223	if ncfilename is None:
3244	3224	filepath = pathlib.Path(cdlfilename)

3259	3239	"""
3260	3240	`tocdl(self, coordvars=False, data=False, outfile=None)`
3261	3241
3262		call `ncdump` via subprocess to create [CDL](https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#cdl_guide)
3263		text representation of Dataset. Requires `ncdump` to be installed and in `$PATH`.
	3242	call [ncdump][ncdump] via subprocess to create [CDL][cdl]
	3243	text representation of Dataset. Requires [ncdump][ncdump]
	3244	to be installed and in `$PATH`.
3264	3245
3265	3246	`coordvars`: include coordinate variable data (via `ncdump -c`). Default False
3266	3247
3267	3248	`data`: if True, write out variable data (Default False).
3268	3249
3269	3250	`outfile`: If not None, file to output ncdump to. Default is to return a string.
	3251
	3252	[ncdump]: https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#ncdump_guide
	3253	[cdl]: https://www.unidata.ucar.edu/software/netcdf/docs/netcdf_utilities_guide.html#cdl_guide
3270	3254	"""
3271	3255	self.sync()
3272	3256	if coordvars:

3287	3271	cdef class Group(Dataset):
3288	3272	"""
3289	3273	Groups define a hierarchical namespace within a netCDF file. They are
3290		analogous to directories in a unix filesystem. Each [Group](#Group) behaves like
3291		a [Dataset](#Dataset) within a Dataset, and can contain it's own variables,
3292		dimensions and attributes (and other Groups). See [Group.__init__](#Group.__init__)
	3274	analogous to directories in a unix filesystem. Each `Group` behaves like
	3275	a `Dataset` within a Dataset, and can contain it's own variables,
	3276	dimensions and attributes (and other Groups). See `Group.__init__`
3293	3277	for more details.
3294	3278
3295		[Group](#Group) inherits from [Dataset](#Dataset), so all the
3296		[Dataset](#Dataset) class methods and variables are available
3297		to a [Group](#Group) instance (except the `close` method).
	3279	`Group` inherits from `Dataset`, so all the
	3280	`Dataset` class methods and variables are available
	3281	to a `Group` instance (except the `close` method).
3298	3282
3299	3283	Additional read-only class variables:
3300	3284

3303	3287	def __init__(self, parent, name, **kwargs):
3304	3288	"""
3305	3289	`__init__(self, parent, name)`
3306		[Group](#Group) constructor.
3307
3308		`parent`: [Group](#Group) instance for the parent group. If being created
3309		in the root group, use a [Dataset](#Dataset) instance.
	3290	`Group` constructor.
	3291
	3292	`parent`: `Group` instance for the parent group. If being created
	3293	in the root group, use a `Dataset` instance.
3310	3294
3311	3295	`name`: - Name of the group.
3312	3296
3313		*Note*: [Group](#Group) instances should be created using the
3314		[Dataset.createGroup](#Dataset.createGroup) method of a [Dataset](#Dataset) instance, or
3315		another [Group](#Group) instance, not using this class directly.
	3297	*Note*: `Group` instances should be created using the
	3298	`Dataset.createGroup` method of a `Dataset` instance, or
	3299	another `Group` instance, not using this class directly.
3316	3300	"""
3317	3301	cdef char *groupname
3318	3302	# flag to indicate that Variables in this Group support orthogonal indexing.

3362	3346	"""
3363	3347	`close(self)`
3364	3348
3365		overrides [Dataset](#Dataset) close method which does not apply to [Group](#Group)
	3349	overrides `Dataset` close method which does not apply to `Group`
3366	3350	instances, raises IOError."""
3367		raise IOError('cannot close a [Group](#Group) (only applies to Dataset)')
	3351	raise IOError('cannot close a `Group` (only applies to Dataset)')
3368	3352
3369	3353
3370	3354	cdef class Dimension:
3371	3355	"""
3372		A netCDF [Dimension](#Dimension) is used to describe the coordinates of a [Variable](#Variable).
3373		See [Dimension.__init__](#Dimension.__init__) for more details.
3374
3375		The current maximum size of a [Dimension](#Dimension) instance can be obtained by
3376		calling the python `len` function on the [Dimension](#Dimension) instance. The
3377		[Dimension.isunlimited](#Dimension.isunlimited) method of a [Dimension](#Dimension) instance can be used to
	3356	A netCDF `Dimension` is used to describe the coordinates of a `Variable`.
	3357	See `Dimension.__init__` for more details.
	3358
	3359	The current maximum size of a `Dimension` instance can be obtained by
	3360	calling the python `len` function on the `Dimension` instance. The
	3361	`Dimension.isunlimited` method of a `Dimension` instance can be used to
3378	3362	determine if the dimension is unlimited.
3379	3363
3380	3364	Read-only class variables:
3381	3365
3382		`name`: String name, used when creating a [Variable](#Variable) with
3383		[Dataset.createVariable](#Dataset.createVariable).
3384
3385		`size`: Current [Dimension](#Dimension) size (same as `len(d)`, where `d` is a
3386		[Dimension](#Dimension) instance).
	3366	`name`: String name, used when creating a `Variable` with
	3367	`Dataset.createVariable`.
	3368
	3369	`size`: Current `Dimension` size (same as `len(d)`, where `d` is a
	3370	`Dimension` instance).
3387	3371	"""
3388	3372	cdef public int _dimid, _grpid
3389	3373	cdef public _data_model, _name, _grp

3392	3376	"""
3393	3377	`__init__(self, group, name, size=None)`
3394	3378
3395		[Dimension](#Dimension) constructor.
3396
3397		`group`: [Group](#Group) instance to associate with dimension.
	3379	`Dimension` constructor.
	3380
	3381	`group`: `Group` instance to associate with dimension.
3398	3382
3399	3383	`name`: Name of the dimension.
3400	3384
3401	3385	`size`: Size of the dimension. `None` or 0 means unlimited. (Default `None`).
3402	3386
3403		*Note*: [Dimension](#Dimension) instances should be created using the
3404		[Dataset.createDimension](#Dataset.createDimension) method of a [Group](#Group) or
3405		[Dataset](#Dataset) instance, not using [Dimension.__init__](#Dimension.__init__) directly.
	3387	*Note*: `Dimension` instances should be created using the
	3388	`Dataset.createDimension` method of a `Group` or
	3389	`Dataset` instance, not using `Dimension.__init__` directly.
3406	3390	"""
3407	3391	cdef int ierr
3408	3392	cdef char *dimname

3468	3452	(type(self), self._name, len(self))
3469	3453
3470	3454	def __len__(self):
3471		# len([Dimension](#Dimension) instance) returns current size of dimension
	3455	# len(`Dimension` instance) returns current size of dimension
3472	3456	cdef int ierr
3473	3457	cdef size_t lengthp
3474	3458	with nogil:

3480	3464	"""
3481	3465	`group(self)`
3482	3466
3483		return the group that this [Dimension](#Dimension) is a member of."""
	3467	return the group that this `Dimension` is a member of."""
3484	3468	return self._grp
3485	3469
3486	3470	def isunlimited(self):
3487	3471	"""
3488	3472	`isunlimited(self)`
3489	3473
3490		returns `True` if the [Dimension](#Dimension) instance is unlimited, `False` otherwise."""
	3474	returns `True` if the `Dimension` instance is unlimited, `False` otherwise."""
3491	3475	cdef int ierr, n, numunlimdims, ndims, nvars, ngatts, xdimid
3492	3476	cdef int *unlimdimids
3493	3477	if self._data_model == 'NETCDF4':

3520	3504
3521	3505	cdef class Variable:
3522	3506	"""
3523		A netCDF [Variable](#Variable) is used to read and write netCDF data. They are
3524		analogous to numpy array objects. See [Variable.__init__](#Variable.__init__) for more
	3507	A netCDF `Variable` is used to read and write netCDF data. They are
	3508	analogous to numpy array objects. See `Variable.__init__` for more
3525	3509	details.
3526	3510
3527	3511	A list of attribute names corresponding to netCDF attributes defined for
3528		the variable can be obtained with the [Variable.ncattrs](#Variable.ncattrs) method. These
	3512	the variable can be obtained with the `Variable.ncattrs` method. These
3529	3513	attributes can be created by assigning to an attribute of the
3530		[Variable](#Variable) instance. A dictionary containing all the netCDF attribute
	3514	`Variable` instance. A dictionary containing all the netCDF attribute
3531	3515	name/value pairs is provided by the `__dict__` attribute of a
3532		[Variable](#Variable) instance.
	3516	`Variable` instance.
3533	3517
3534	3518	The following class variables are read-only:
3535	3519

3546	3530	`scale`: If True, `scale_factor` and `add_offset` are
3547	3531	applied, and signed integer data is automatically converted to
3548	3532	unsigned integer data if the `_Unsigned` attribute is set.
3549		Default is `True`, can be reset using [Variable.set_auto_scale](#Variable.set_auto_scale) and
3550		[Variable.set_auto_maskandscale](#Variable.set_auto_maskandscale) methods.
	3533	Default is `True`, can be reset using `Variable.set_auto_scale` and
	3534	`Variable.set_auto_maskandscale` methods.
3551	3535
3552	3536	`mask`: If True, data is automatically converted to/from masked
3553	3537	arrays when missing values or fill values are present. Default is `True`, can be
3554		reset using [Variable.set_auto_mask](#Variable.set_auto_mask) and [Variable.set_auto_maskandscale](#Variable.set_auto_maskandscale)
	3538	reset using `Variable.set_auto_mask` and `Variable.set_auto_maskandscale`
3555	3539	methods.
3556	3540
3557	3541	`chartostring`: If True, data is automatically converted to/from character
3558	3542	arrays to string arrays when the `_Encoding` variable attribute is set.
3559	3543	Default is `True`, can be reset using
3560		[Variable.set_auto_chartostring](#Variable.set_auto_chartostring) method.
	3544	`Variable.set_auto_chartostring` method.
3561	3545
3562	3546	`least_significant_digit`: Describes the power of ten of the
3563	3547	smallest decimal place in the data the contains a reliable value. Data is
3564		truncated to this decimal place when it is assigned to the [Variable](#Variable)
	3548	truncated to this decimal place when it is assigned to the `Variable`
3565	3549	instance. If `None`, the data is not truncated.
3566	3550
3567	3551	`__orthogonal_indexing__`: Always `True`. Indicates to client code

3591	3575	chunksizes=None, endian='native',
3592	3576	least_significant_digit=None,fill_value=None,chunk_cache=None)`**
3593	3577
3594		[Variable](#Variable) constructor.
3595
3596		`group`: [Group](#Group) or [Dataset](#Dataset) instance to associate with variable.
	3578	`Variable` constructor.
	3579
	3580	`group`: `Group` or `Dataset` instance to associate with variable.
3597	3581
3598	3582	`name`: Name of the variable.
3599	3583
3600		`datatype`: [Variable](#Variable) data type. Can be specified by providing a
	3584	`datatype`: `Variable` data type. Can be specified by providing a
3601	3585	numpy dtype object, or a string that describes a numpy dtype object.
3602	3586	Supported values, corresponding to `str` attribute of numpy dtype
3603	3587	objects, include `'f4'` (32-bit floating point), `'f8'` (64-bit floating

3610	3594	typecodes can also be used (`'f'` instead of `'f4'`, `'d'` instead of
3611	3595	`'f8'`, `'h'` or `'s'` instead of `'i2'`, `'b'` or `'B'` instead of
3612	3596	`'i1'`, `'c'` instead of `'S1'`, and `'i'` or `'l'` instead of
3613		`'i4'`). `datatype` can also be a [CompoundType](#CompoundType) instance
3614		(for a structured, or compound array), a [VLType](#VLType) instance
	3597	`'i4'`). `datatype` can also be a `CompoundType` instance
	3598	(for a structured, or compound array), a `VLType` instance
3615	3599	(for a variable-length array), or the python `str` builtin
3616	3600	(for a variable-length string array). Numpy string and unicode datatypes with
3617	3601	length greater than one are aliases for `str`.

3620	3604	(defined previously with `createDimension`). Default is an empty tuple
3621	3605	which means the variable is a scalar (and therefore has no dimensions).
3622	3606
3623		`zlib`: if `True`, data assigned to the [Variable](#Variable)
	3607	`zlib`: if `True`, data assigned to the `Variable`
3624	3608	instance is compressed on disk. Default `False`.
3625	3609
3626	3610	`complevel`: the level of zlib compression to use (1 is the fastest,

3669	3653	value that the variable gets filled with before any data is written to it)
3670	3654	is replaced with this value. If fill_value is set to `False`, then
3671	3655	the variable is not pre-filled. The default netCDF fill values can be found
3672		in [default_fillvals](#default_fillvals).
	3656	in the dictionary `netCDF4.default_fillvals`.
3673	3657
3674	3658	`chunk_cache`: If specified, sets the chunk cache size for this variable.
3675		Persists as long as Dataset is open. Use [set_var_chunk_cache](#set_var_chunk_cache) to
	3659	Persists as long as Dataset is open. Use `set_var_chunk_cache` to
3676	3660	change it when Dataset is re-opened.
3677	3661
3678		*Note*: [Variable](#Variable) instances should be created using the
3679		[Dataset.createVariable](#Dataset.createVariable) method of a [Dataset](#Dataset) or
3680		[Group](#Group) instance, not using this class directly.
	3662	*Note*: `Variable` instances should be created using the
	3663	`Dataset.createVariable` method of a `Dataset` or
	3664	`Group` instance, not using this class directly.
3681	3665	"""
3682	3666	cdef int ierr, ndims, icontiguous, ideflate_level, numdims, _grpid
3683	3667	cdef char namstring[NC_MAX_NAME+1]

4106	4090	"""
4107	4091	`group(self)`
4108	4092
4109		return the group that this [Variable](#Variable) is a member of."""
	4093	return the group that this `Variable` is a member of."""
4110	4094	return self._grp
4111	4095
4112	4096	def ncattrs(self):
4113	4097	"""
4114	4098	`ncattrs(self)`
4115	4099
4116		return netCDF attribute names for this [Variable](#Variable) in a list."""
	4100	return netCDF attribute names for this `Variable` in a list."""
4117	4101	return _get_att_names(self._grpid, self._varid)
4118	4102
4119	4103	def setncattr(self,name,value):

4374	4358	"""
4375	4359	`renameAttribute(self, oldname, newname)`
4376	4360
4377		rename a [Variable](#Variable) attribute named `oldname` to `newname`."""
	4361	rename a `Variable` attribute named `oldname` to `newname`."""
4378	4362	cdef int ierr
4379	4363	cdef char *oldnamec
4380	4364	cdef char *newnamec

4597	4581	# type, signed or unsigned, because the byte ranges are too
4598	4582	# small to assume one of the values should appear as a missing
4599	4583	# value unless a _FillValue attribute is set explicitly."
4600		if no_fill != 1 or self.dtype.str[1:] not in ['u1','i1']:
	4584	# (do this only for non-vlens, since vlens don't have a default _FillValue)
	4585	if not self._isvlen and (no_fill != 1 or self.dtype.str[1:] not in ['u1','i1']):
4601	4586	fillval = numpy.array(default_fillvals[self.dtype.str[1:]],self.dtype)
4602	4587	has_fillval = data == fillval
4603	4588	# if data is an array scalar, has_fillval will be a boolean.

4895	4880	# if auto scaling is to be done, don't cast to an integer yet.
4896	4881	if self.scale and self.dtype.kind in 'iu' and \
4897	4882	hasattr(self, 'scale_factor') or hasattr(self, 'add_offset'):
4898		data = numpy.array(data,numpy.float_)
	4883	data = numpy.array(data,numpy.float64)
4899	4884	else:
4900	4885	data = numpy.array(data,self.dtype)
4901	4886

5037	5022	missing_value variable attribute. The fill_value of the masked array
5038	5023	is set to the missing_value attribute (if it exists), otherwise
5039	5024	the netCDF _FillValue attribute (which has a default value
5040		for each data type). When data is written to a variable, the masked
	5025	for each data type). If the variable has no missing_value attribute, the
	5026	_FillValue is used instead. If the variable has valid_min/valid_max and
	5027	missing_value attributes, data outside the specified range will be masked.
	5028	When data is written to a variable, the masked
5041	5029	array is converted back to a regular numpy array by replacing all the
5042	5030	masked values by the missing_value attribute of the variable (if it
5043	5031	exists). If the variable has no missing_value attribute, the _FillValue
5044		is used instead. If the variable has valid_min/valid_max and
5045		missing_value attributes, data outside the specified range will be
5046		set to missing_value.
	5032	is used instead.
5047	5033
5048	5034	If `maskandscale` is set to `True`, and the variable has a
5049	5035	`scale_factor` or an `add_offset` attribute, then data read

5126	5112	missing_value variable attribute. The fill_value of the masked array
5127	5113	is set to the missing_value attribute (if it exists), otherwise
5128	5114	the netCDF _FillValue attribute (which has a default value
5129		for each data type). When data is written to a variable, the masked
	5115	for each data type). If the variable has no missing_value attribute, the
	5116	_FillValue is used instead. If the variable has valid_min/valid_max and
	5117	missing_value attributes, data outside the specified range will be masked.
	5118	When data is written to a variable, the masked
5130	5119	array is converted back to a regular numpy array by replacing all the
5131	5120	masked values by the missing_value attribute of the variable (if it
5132	5121	exists). If the variable has no missing_value attribute, the _FillValue
5133		is used instead. If the variable has valid_min/valid_max and
5134		missing_value attributes, data outside the specified range will be
5135		set to missing_value.
	5122	is used instead.
5136	5123
5137	5124	The default value of `mask` is `True`
5138	5125	(automatic conversions are performed).

5477	5464	"""
5478	5465	`get_dims(self)`
5479	5466
5480		return a tuple of [Dimension](#Dimension) instances associated with this
5481		[Variable](#Variable).
	5467	return a tuple of `Dimension` instances associated with this
	5468	`Variable`.
5482	5469	"""
5483	5470	return tuple(_find_dim(self._grp, dim) for dim in self.dimensions)
5484	5471

5490	5477
5491	5478	cdef class CompoundType:
5492	5479	"""
5493		A [CompoundType](#CompoundType) instance is used to describe a compound data
5494		type, and can be passed to the the [Dataset.createVariable](#Dataset.createVariable) method of
5495		a [Dataset](#Dataset) or [Group](#Group) instance.
	5480	A `CompoundType` instance is used to describe a compound data
	5481	type, and can be passed to the the `Dataset.createVariable` method of
	5482	a `Dataset` or `Group` instance.
5496	5483	Compound data types map to numpy structured arrays.
5497		See [CompoundType.__init__](#CompoundType.__init__) for more details.
	5484	See `CompoundType.__init__` for more details.
5498	5485
5499	5486	The instance variables `dtype` and `name` should not be modified by
5500	5487	the user.

5507	5494
5508	5495	CompoundType constructor.
5509	5496
5510		`group`: [Group](#Group) instance to associate with the compound datatype.
	5497	`group`: `Group` instance to associate with the compound datatype.
5511	5498
5512	5499	`datatype`: A numpy dtype object describing a structured (a.k.a record)
5513	5500	array. Can be composed of homogeneous numeric or character data types, or

5521	5508	instances (so create CompoundType instances for the innermost structures
5522	5509	first).
5523	5510
5524		*Note 2*: [CompoundType](#CompoundType) instances should be created using the
5525		[Dataset.createCompoundType](#Dataset.createCompoundType)
5526		method of a [Dataset](#Dataset) or [Group](#Group) instance, not using this class directly.
	5511	*Note 2*: `CompoundType` instances should be created using the
	5512	`Dataset.createCompoundType` method of a `Dataset` or
	5513	`Group` instance, not using this class directly.
5527	5514	"""
5528	5515	cdef nc_type xtype
5529	5516	# convert dt to a numpy datatype object

5795	5782
5796	5783	cdef class VLType:
5797	5784	"""
5798		A [VLType](#VLType) instance is used to describe a variable length (VLEN) data
5799		type, and can be passed to the the [Dataset.createVariable](#Dataset.createVariable) method of
5800		a [Dataset](#Dataset) or [Group](#Group) instance. See
5801		[VLType.__init__](#VLType.__init__)for more details.
	5785	A `VLType` instance is used to describe a variable length (VLEN) data
	5786	type, and can be passed to the the `Dataset.createVariable` method of
	5787	a `Dataset` or `Group` instance. See
	5788	`VLType.__init__` for more details.
5802	5789
5803	5790	The instance variables `dtype` and `name` should not be modified by
5804	5791	the user.

5811	5798
5812	5799	VLType constructor.
5813	5800
5814		`group`: [Group](#Group) instance to associate with the VLEN datatype.
	5801	`group`: `Group` instance to associate with the VLEN datatype.
5815	5802
5816	5803	`datatype`: An numpy dtype object describing the component type for the
5817	5804	variable length array.

5819	5806	`datatype_name`: a Python string containing a description of the
5820	5807	VLEN data type.
5821	5808
5822		*`Note`*: [VLType](#VLType) instances should be created using the
5823		[Dataset.createVLType](#Dataset.createVLType)
5824		method of a [Dataset](#Dataset) or [Group](#Group) instance, not using this class directly.
	5809	*`Note`*: `VLType` instances should be created using the
	5810	`Dataset.createVLType` method of a `Dataset` or
	5811	`Group` instance, not using this class directly.
5825	5812	"""
5826	5813	cdef nc_type xtype
5827	5814	if 'typeid' in kwargs:

5904	5891
5905	5892	cdef class EnumType:
5906	5893	"""
5907		A [EnumType](#EnumType) instance is used to describe an Enum data
5908		type, and can be passed to the the [Dataset.createVariable](#Dataset.createVariable) method of
5909		a [Dataset](#Dataset) or [Group](#Group) instance. See
5910		[EnumType.__init__](#EnumType.__init__) for more details.
	5894	A `EnumType` instance is used to describe an Enum data
	5895	type, and can be passed to the the `Dataset.createVariable` method of
	5896	a `Dataset` or `Group` instance. See
	5897	`EnumType.__init__` for more details.
5911	5898
5912	5899	The instance variables `dtype`, `name` and `enum_dict` should not be modified by
5913	5900	the user.

5920	5907
5921	5908	EnumType constructor.
5922	5909
5923		`group`: [Group](#Group) instance to associate with the VLEN datatype.
	5910	`group`: `Group` instance to associate with the VLEN datatype.
5924	5911
5925	5912	`datatype`: An numpy integer dtype object describing the base type
5926	5913	for the Enum.

5931	5918	`enum_dict`: a Python dictionary containing the Enum field/value
5932	5919	pairs.
5933	5920
5934		*`Note`*: [EnumType](#EnumType) instances should be created using the
5935		[Dataset.createEnumType](#Dataset.createEnumType)
5936		method of a [Dataset](#Dataset) or [Group](#Group) instance, not using this class directly.
	5921	*`Note`*: `EnumType` instances should be created using the
	5922	`Dataset.createEnumType` method of a `Dataset` or
	5923	`Group` instance, not using this class directly.
5937	5924	"""
5938	5925	cdef nc_type xtype
5939	5926	if 'typeid' in kwargs:

6122	6109
6123	6110	Adapted from [pycdf](http://pysclint.sourceforge.net/pycdf) by Andre Gosselin.
6124	6111
6125		Example usage (See [MFDataset.__init__](#MFDataset.__init__) for more details):
	6112	Example usage (See `MFDataset.__init__` for more details):
6126	6113
6127	6114	```python
6128	6115	>>> import numpy as np

6585	6572	Class providing an interface to a MFDataset time Variable by imposing a unique common
6586	6573	time unit and/or calendar to all files.
6587	6574
6588		Example usage (See [MFTime.__init__](#MFTime.__init__) for more details):
	6575	Example usage (See `MFTime.__init__` for more details):
6589	6576
6590	6577	```python
6591		>>> import numpy
	6578	>>> import numpy as np
6592	6579	>>> f1 = Dataset("mftest_1.nc","w", format="NETCDF4_CLASSIC")
6593	6580	>>> f2 = Dataset("mftest_2.nc","w", format="NETCDF4_CLASSIC")
6594	6581	>>> f1.createDimension("time",None)

6599	6586	>>> t2.units = "days since 2000-02-01"
6600	6587	>>> t1.calendar = "standard"
6601	6588	>>> t2.calendar = "standard"
6602		>>> t1[:] = numpy.arange(31)
6603		>>> t2[:] = numpy.arange(30)
	6589	>>> t1[:] = np.arange(31)
	6590	>>> t2[:] = np.arange(30)
6604	6591	>>> f1.close()
6605	6592	>>> f2.close()
6606	6593	>>> # Read the two files in at once, in one Dataset.

6623	6610	Create a time Variable with units consistent across a multifile
6624	6611	dataset.
6625	6612
6626		`time`: Time variable from a [MFDataset](#MFDataset).
	6613	`time`: Time variable from a `MFDataset`.
6627	6614
6628	6615	`units`: Time units, for example, `'days since 1979-01-01'`. If `None`,
6629	6616	use the units from the master variable.

+24

-0

src/netCDF4.egg-info/PKG-INFO less more

	0	Metadata-Version: 1.1
	1	Name: netCDF4
	2	Version: 1.5.7
	3	Summary: Provides an object-oriented python interface to the netCDF version 4 library.
	4	Home-page: http://github.com/Unidata/netcdf4-python
	5	Author: Jeff Whitaker
	6	Author-email: jeffrey.s.whitaker@noaa.gov
	7	License: License :: OSI Approved :: MIT License
	8	Download-URL: http://python.org/pypi/netCDF4
	9	Description: netCDF version 4 has many features not found in earlier versions of the library, such as hierarchical groups, zlib compression, multiple unlimited dimensions, and new data types. It is implemented on top of HDF5. This module implements most of the new features, and can read and write netCDF files compatible with older versions of the library. The API is modelled after Scientific.IO.NetCDF, and should be familiar to users of that module.
	10
	11	This project is hosted on a `GitHub repository <https://github.com/Unidata/netcdf4-python>`_ where you may access the most up-to-date source.
	12	Keywords: numpy,netcdf,data,science,network,oceanography,meteorology,climate
	13	Platform: any
	14	Classifier: Development Status :: 3 - Alpha
	15	Classifier: Programming Language :: Python :: 3
	16	Classifier: Programming Language :: Python :: 3.6
	17	Classifier: Programming Language :: Python :: 3.7
	18	Classifier: Programming Language :: Python :: 3.8
	19	Classifier: Intended Audience :: Science/Research
	20	Classifier: License :: OSI Approved :: MIT License
	21	Classifier: Topic :: Software Development :: Libraries :: Python Modules
	22	Classifier: Topic :: System :: Archiving :: Compression
	23	Classifier: Operating System :: OS Independent

+93

-0

src/netCDF4.egg-info/SOURCES.txt less more

	0	Changelog
	1	LICENSE
	2	MANIFEST.in
	3	README.htmldocs
	4	README.md
	5	README.release
	6	README.wheels.md
	7	checkversion.py
	8	create_docs.sh
	9	setup.cfg
	10	setup.py
	11	docs/index.html
	12	examples/README.md
	13	examples/bench.py
	14	examples/bench_compress.py
	15	examples/bench_compress2.py
	16	examples/bench_compress3.py
	17	examples/bench_diskless.py
	18	examples/json_att.py
	19	examples/mpi_example.py
	20	examples/mpi_example_compressed.py
	21	examples/subset.py
	22	examples/test_stringarr.py
	23	examples/threaded_read.py
	24	examples/tutorial.py
	25	include/membuf.pyx
	26	include/mpi-compat.h
	27	include/netCDF4.pxi
	28	man/nc3tonc4.1
	29	man/nc4tonc3.1
	30	man/ncinfo.1
	31	src/netCDF4/__init__.py
	32	src/netCDF4/_netCDF4.pyx
	33	src/netCDF4/utils.py
	34	src/netCDF4.egg-info/PKG-INFO
	35	src/netCDF4.egg-info/SOURCES.txt
	36	src/netCDF4.egg-info/dependency_links.txt
	37	src/netCDF4.egg-info/entry_points.txt
	38	src/netCDF4.egg-info/requires.txt
	39	src/netCDF4.egg-info/top_level.txt
	40	test/20171025_2056.Cloud_Top_Height.nc
	41	test/CRM032_test1.nc
	42	test/issue671.nc
	43	test/issue672.nc
	44	test/netcdf_dummy_file.nc
	45	test/run_all.py
	46	test/test_gold.nc
	47	test/tst_Unsigned.py
	48	test/tst_atts.py
	49	test/tst_cdf5.py
	50	test/tst_cdl.py
	51	test/tst_chunk_cache.py
	52	test/tst_compound_alignment.py
	53	test/tst_compoundatt.py
	54	test/tst_compoundvar.py
	55	test/tst_compression.py
	56	test/tst_create_mem.py
	57	test/tst_dap.py
	58	test/tst_dims.py
	59	test/tst_diskless.py
	60	test/tst_endian.py
	61	test/tst_enum.py
	62	test/tst_fancyslicing.py
	63	test/tst_filepath.py
	64	test/tst_get_variables_by_attributes.py
	65	test/tst_grps.py
	66	test/tst_grps2.py
	67	test/tst_issue908.py
	68	test/tst_masked.py
	69	test/tst_masked2.py
	70	test/tst_masked3.py
	71	test/tst_masked4.py
	72	test/tst_masked5.py
	73	test/tst_masked6.py
	74	test/tst_multifile.py
	75	test/tst_multifile2.py
	76	test/tst_open_mem.py
	77	test/tst_refcount.py
	78	test/tst_rename.py
	79	test/tst_scalarvar.py
	80	test/tst_scaled.py
	81	test/tst_shape.py
	82	test/tst_slicing.py
	83	test/tst_stringarr.py
	84	test/tst_types.py
	85	test/tst_unicode.py
	86	test/tst_unicode3.py
	87	test/tst_unicodeatt.py
	88	test/tst_unlimdim.py
	89	test/tst_utils.py
	90	test/tst_vars.py
	91	test/tst_vlen.py
	92	test/ubyte.nc⏎

-0

src/netCDF4.egg-info/dependency_links.txt less more

-0

src/netCDF4.egg-info/entry_points.txt less more

	0	[console_scripts]
	1	nc3tonc4 = netCDF4.utils:nc3tonc4
	2	nc4tonc3 = netCDF4.utils:nc4tonc3
	3	ncinfo = netCDF4.utils:ncinfo
	4

-0

src/netCDF4.egg-info/requires.txt less more

	0	cftime
	1	numpy>=1.9

-0

src/netCDF4.egg-info/top_level.txt less more

netCDF4

+12

-22

test/tst_atts.py less more

5	5	import tempfile
6	6	import warnings
7	7
8		import numpy as NP
	8	import numpy as np
9	9	from collections import OrderedDict
10	10	from numpy.random.mtrand import uniform
11	11
12	12	import netCDF4
13	13
14	14	# test attribute creation.
15		FILE_NAME = tempfile.NamedTemporaryFile(suffix='.nc', delete=False).name
	15	#FILE_NAME = tempfile.NamedTemporaryFile(suffix='.nc', delete=False).name
	16	FILE_NAME = 'tst_atts.nc'
16	17	VAR_NAME="dummy_var"
17	18	GROUP_NAME = "dummy_group"
18	19	DIM1_NAME="x"

25	26	EMPTYSTRATT = ''
26	27	INTATT = 1
27	28	FLOATATT = math.pi
28		SEQATT = NP.arange(10)
	29	SEQATT = np.arange(10)
29	30	STRINGSEQATT = ['mary ','','had ','a ','little ','lamb',]
30	31	#ATTDICT = {'stratt':STRATT,'floatatt':FLOATATT,'seqatt':SEQATT,
31	32	# 'stringseqatt':''.join(STRINGSEQATT), # changed in issue #770

105	106	raise ValueError('This test should have failed.')
106	107	# issue #485 (triggers segfault in C lib
107	108	# with version 1.2.1 without pull request #486)
108		f.foo = NP.array('bar','S')
109		f.foo = NP.array('bar','U')
	109	f.foo = np.array('bar','S')
	110	f.foo = np.array('bar','U')
110	111	# issue #529 write string attribute as NC_CHAR unless
111	112	# it can't be decoded to ascii. Add setncattr_string
112	113	# method to force NC_STRING.

144	145	# global attributes.
145	146	# check __dict__ method for accessing all netCDF attributes.
146	147	for key,val in ATTDICT.items():
147		if type(val) == NP.ndarray:
	148	if type(val) == np.ndarray:
148	149	assert f.__dict__[key].tolist() == val.tolist()
149	150	else:
150	151	assert f.__dict__[key] == val

161	162	# variable attributes.
162	163	# check __dict__ method for accessing all netCDF attributes.
163	164	for key,val in ATTDICT.items():
164		if type(val) == NP.ndarray:
	165	if type(val) == np.ndarray:
165	166	assert v.__dict__[key].tolist() == val.tolist()
166	167	else:
167	168	assert v.__dict__[key] == val

177	178	assert v.getncattr('foo') == 1
178	179	assert v.getncattr('bar') == 2
179	180	# check type of attributes using ncdump (issue #529)
180		try: # ncdump may not be on the system PATH
181		nc_proc = subprocess.Popen(
182		['ncdump', '-h', FILE_NAME], stdout=subprocess.PIPE)
183		except OSError:
184		warnings.warn('"ncdump" not on system path; cannot test '
185		'read of some attributes')
186		pass
187		else: # We do have ncdump output
188		dep = nc_proc.communicate()[0]
189		try: # python 2
190		ncdump_output = dep.split('\n')
191		except TypeError: # python 3
192		ncdump_output = str(dep,encoding='utf-8').split('\n')
	181	if not os.getenv('NO_CDL'):
	182	ncdump_output = f.tocdl()
193	183	for line in ncdump_output:
194	184	line = line.strip('\t\n\r')
195	185	line = line.strip()# Must be done another time for group variables

201	191	# check attributes in subgroup.
202	192	# global attributes.
203	193	for key,val in ATTDICT.items():
204		if type(val) == NP.ndarray:
	194	if type(val) == np.ndarray:
205	195	assert g.__dict__[key].tolist() == val.tolist()
206	196	else:
207	197	assert g.__dict__[key] == val

214	204	assert g.stringseqatt == STRINGSEQATT
215	205	assert g.stringseqatt_array == STRINGSEQATT
216	206	for key,val in ATTDICT.items():
217		if type(val) == NP.ndarray:
	207	if type(val) == np.ndarray:
218	208	assert v1.__dict__[key].tolist() == val.tolist()
219	209	else:
220	210	assert v1.__dict__[key] == val

-1

test/tst_dims.py less more

1	1	import unittest
2	2	import os
3	3	import tempfile
4		import numpy as NP
5	4	from numpy.random.mtrand import uniform
6	5	import netCDF4
7	6

-1

test/tst_grps.py less more

1	1	import unittest
2	2	import os
3	3	import tempfile
4		import numpy as NP
5	4	import netCDF4
6	5
7	6	# test group creation.

-1

test/tst_grps2.py less more

1	1	import unittest
2	2	import os
3	3	import tempfile
4		import numpy as NP
5	4	import netCDF4
6	5
7	6	# test implicit group creation by using unix-like paths

+13

-13

test/tst_masked.py less more

1	1	import unittest
2	2	import os
3	3	import tempfile
4		import numpy as NP
	4	import numpy as np
5	5	from numpy import ma
6	6	from numpy.testing import assert_array_equal, assert_array_almost_equal
7	7	from numpy.random.mtrand import uniform

18	18	ranarr = 100.*uniform(size=(ndim))
19	19	ranarr2 = 100.*uniform(size=(ndim))
20	20	# used for checking vector missing_values
21		arr3 = NP.linspace(0,9,ndim)
22		mask = NP.zeros(ndim,NP.bool); mask[-1]=True; mask[-2]=True
23		marr3 = NP.ma.array(arr3, mask=mask, dtype=NP.int32)
	21	arr3 = np.linspace(0,9,ndim)
	22	mask = np.zeros(ndim,np.bool_); mask[-1]=True; mask[-2]=True
	23	marr3 = np.ma.array(arr3, mask=mask, dtype=np.int32)
24	24	packeddata = 10.*uniform(size=(ndim))
25	25	missing_value = -9999.
26		missing_value2 = NP.nan
	26	missing_value2 = np.nan
27	27	missing_value3 = [8,9]
28	28	ranarr[::2] = missing_value
29	29	ranarr2[::2] = missing_value2
30		NP.seterr(invalid='ignore') # silence warnings from ma.masked_values
	30	np.seterr(invalid='ignore') # silence warnings from ma.masked_values
31	31	maskedarr = ma.masked_values(ranarr,missing_value)
32	32	#maskedarr2 = ma.masked_values(ranarr2,missing_value2)
33	33	maskedarr2 = ma.masked_invalid(ranarr2)
34	34	scale_factor = (packeddata.max()-packeddata.min())/(2.*32766.)
35	35	add_offset = 0.5*(packeddata.max()+packeddata.min())
36		packeddata2 = NP.around((packeddata-add_offset)/scale_factor).astype('i2')
	36	packeddata2 = np.around((packeddata-add_offset)/scale_factor).astype('i2')
37	37
38	38	class PrimitiveTypesTestCase(unittest.TestCase):
39	39

70	70	doh2[0] = 1.
71	71	# added test for issue 515
72	72	file.createDimension('x',1)
73		v = file.createVariable('v',NP.float,'x',fill_value=-9999)
	73	v = file.createVariable('v',np.float64,'x',fill_value=-9999)
74	74	file.close()
75	75
76	76	# issue #972: when auto_fill off byte arrays (u1,i1) should

118	118	assert_array_almost_equal(datamasked[:].filled(),ranarr)
119	119	assert_array_almost_equal(datamasked2[:].filled(),ranarr2)
120	120	assert_array_almost_equal(datapacked[:],packeddata,decimal=4)
121		assert(datapacked3[:].dtype == NP.float)
	121	assert(datapacked3[:].dtype == np.float64)
122	122	# added to test fix to issue 46 (result before r865 was 10)
123	123	assert_array_equal(datapacked2[0],11)
124	124	# added test for issue 515
125		assert(file['v'][0] is NP.ma.masked)
	125	assert(file['v'][0] is np.ma.masked)
126	126	file.close()
127	127	# issue 766
128		NP.seterr(invalid='raise')
	128	np.seterr(invalid='raise')
129	129	f = netCDF4.Dataset(self.file, 'w')
130	130	f.createDimension('dimension', 2)
131		f.createVariable('variable', NP.float32, dimensions=('dimension',))
132		f['variable'][:] = NP.nan
	131	f.createVariable('variable', np.float32, dimensions=('dimension',))
	132	f['variable'][:] = np.nan
133	133	data = f['variable'][:] # should not raise an error
134	134	f.close()
135	135	# issue #972

-3

test/tst_masked2.py less more

15	15	FILE_NAME2 = tempfile.NamedTemporaryFile(suffix='.nc', delete=False).name
16	16	FILE_NAME3 = tempfile.NamedTemporaryFile(suffix='.nc', delete=False).name
17	17	datacheck1 =\
18		ma.array([0,5000.0,4000.0,0],dtype=np.float,mask=[True,False,False,True])
	18	ma.array([0,5000.0,4000.0,0],dtype=np.float64,mask=[True,False,False,True])
19	19	datacheck2 =\
20		ma.array([3000.0,5000.0,4000.0,0],dtype=np.float,mask=[False,False,False,True])
	20	ma.array([3000.0,5000.0,4000.0,0],dtype=np.float64,mask=[False,False,False,True])
21	21	datacheck3 =\
22		ma.array([3000.0,5000.0,0,2000.0],dtype=np.float,mask=[False,False,True,False])
	22	ma.array([3000.0,5000.0,0,2000.0],dtype=np.float64,mask=[False,False,True,False])
23	23	mask = [False,True,False,False]
24	24	datacheck4 = ma.array([1.5625,0,3.75,4.125],mask=mask,dtype=np.float32)
25	25	fillval = default_fillvals[datacheck4.dtype.str[1:]]

-1

test/tst_rename.py less more

1	1	import unittest
2	2	import os
3	3	import tempfile
4		import numpy as NP
5	4	import netCDF4
6	5	from netCDF4 import __has_rename_grp__
7	6

-1

test/tst_scalarvar.py less more

1	1	import unittest
2	2	import os
3	3	import tempfile
4		import numpy as NP
5	4	from numpy.testing import assert_almost_equal
6	5	import netCDF4
7	6	import math

-2

test/tst_shape.py less more

4	4	file_name = tempfile.NamedTemporaryFile(suffix='.nc', delete=False).name
5	5	xdim=None; ydim=121; zdim=169
6	6	datashape = (ydim,zdim)
7		data = np.ones(datashape,dtype=np.float_)
	7	data = np.ones(datashape,dtype=np.float64)
8	8
9	9	class ShapeTestCase(unittest.TestCase):
10	10

14	14	f.createDimension('x',xdim)
15	15	f.createDimension('y',ydim)
16	16	f.createDimension('z',zdim)
17		v = f.createVariable('data',np.float_,('x','y','z'))
	17	v = f.createVariable('data',np.float64,('x','y','z'))
18	18	f.close()
19	19
20	20	def tearDown(self):

+13

-13

test/tst_slicing.py less more

128	128	nlons = 12; lon = f.createDimension('lon',nlons)
129	129	nlevs = 1; lev = f.createDimension('lev',nlevs)
130	130	time = f.createDimension('time',None)
131		var = f.createVariable('var',np.float_,('time','lev','lat','lon'))
	131	var = f.createVariable('var',np.float64,('time','lev','lat','lon'))
132	132	a = np.random.uniform(size=(10,nlevs,nlats,nlons))
133	133	var[0:10] = a
134	134	f.close()

167	167	nlats = 11; lat = f.createDimension('lat',nlats)
168	168	nlons = 20; lon = f.createDimension('lon',nlons)
169	169	time = f.createDimension('time',None)
170		var = f.createVariable('var',np.float_,('time','lat','lon'))
	170	var = f.createVariable('var',np.float64,('time','lat','lon'))
171	171	a = np.random.uniform(size=(3,nlats,nlons))
172	172	var[[True,True,False,False,False,True]] = a
173	173	var[0,2.0,"-1"] = 0 # issue 312

202	202	td = nc.createDimension('t',None)
203	203	xd = nc.createDimension('x',33)
204	204	yd = nc.createDimension('y',4)
205		v = nc.createVariable('v',np.float_,('t','x','y'))
	205	v = nc.createVariable('v',np.float64,('t','x','y'))
206	206	nc.close()
207	207	nc = Dataset(self.file)
208	208	data = np.empty(nc['v'].shape, nc['v'].dtype)

215	215	f.createDimension('d1',3)
216	216	f.createDimension('d2',None)
217	217	f.createDimension('d3',5)
218		f.createVariable('v2',np.float_,('d1','d2','d3'))
	218	f.createVariable('v2',np.float64,('d1','d2','d3'))
219	219	f['v2'][:] = np.zeros((3,4,5))
220	220	f['v2'][0,:,0] = np.arange(4)
221	221	f['v2'][0,:,:] = np.ones((4,5))

226	226	f.createDimension('time',2)
227	227	f.createDimension('lat',10)
228	228	f.createDimension('lon',9)
229		f.createVariable('v1',np.int,('time', 'lon','lat',))
	229	f.createVariable('v1',np.int64,('time', 'lon','lat',))
230	230	arr = np.arange(9*10).reshape((9, 10))
231	231	f['v1'][:] = arr
232	232	assert_array_equal(f['v1'][:],np.broadcast_to(arr,f['v1'].shape))

238	238	with Dataset(self.file,'w') as f:
239	239	f.createDimension('d1',3)
240	240	f.createDimension('d2',None)
241		f.createVariable('v1',np.int,('d2','d1',))
242		f['v1'][0] = np.arange(3,dtype=np.int)
243		f['v1'][1:3] = np.arange(3,dtype=np.int)
	241	f.createVariable('v1',np.int64,('d2','d1',))
	242	f['v1'][0] = np.arange(3,dtype=np.int64)
	243	f['v1'][1:3] = np.arange(3,dtype=np.int64)
244	244	assert_array_equal(f['v1'][:], np.broadcast_to(np.arange(3),(3,3)))
245		f.createVariable('v2',np.int,('d1','d2',))
246		f['v2'][:,0] = np.arange(3,dtype=np.int)
247		f['v2'][:,1:3] = np.arange(6,dtype=np.int).reshape(3,2)
248		assert_array_equal(f['v2'][:,1:3],np.arange(6,dtype=np.int).reshape(3,2))
249		assert_array_equal(f['v2'][:,0],np.arange(3,dtype=np.int))
	245	f.createVariable('v2',np.int64,('d1','d2',))
	246	f['v2'][:,0] = np.arange(3,dtype=np.int64)
	247	f['v2'][:,1:3] = np.arange(6,dtype=np.int64).reshape(3,2)
	248	assert_array_equal(f['v2'][:,1:3],np.arange(6,dtype=np.int64).reshape(3,2))
	249	assert_array_equal(f['v2'][:,0],np.arange(3,dtype=np.int64))
250	250
251	251	def test_issue1083(self):
252	252	with Dataset(self.file, "w") as nc:

-1

test/tst_types.py less more

85	85	else:
86	86	assert(v2._FillValue == u'') # python 2
87	87	assert(str(issue273_data) == str(v2[:]))
88		# isse 707 (don't apply missing_value if cast to variable type is
	88	# issue 707 (don't apply missing_value if cast to variable type is
89	89	# unsafe)
90	90	v3 = f.variables['issue707']
91	91	assert_array_equal(v3[:],-1*np.ones(n2dim,v3.dtype))

-1

test/tst_unicode.py less more

17	17	f.attribute2 = ATT2
18	18	f.attribute3 = ATT3
19	19	d = f.createDimension(DIM_NAME, None)
20		v = f.createVariable(VAR_NAME, np.float, (DIM_NAME,))
	20	v = f.createVariable(VAR_NAME, np.float64, (DIM_NAME,))
21	21	f.close()
22	22
23	23	def tearDown(self):

-1

test/tst_unicode3.py less more

17	17	f.attribute2 = ATT2
18	18	f.attribute3 = ATT3
19	19	d = f.createDimension(DIM_NAME, None)
20		v = f.createVariable(VAR_NAME, np.float_, (DIM_NAME,))
	20	v = f.createVariable(VAR_NAME, np.float64, (DIM_NAME,))
21	21	f.close()
22	22
23	23	def tearDown(self):

-2

test/tst_unlimdim.py less more

1	1	import unittest
2	2	import os
3	3	import tempfile
4		import numpy as NP
	4	import numpy as np
5	5	from numpy.random.mtrand import uniform
6	6	from numpy.testing import assert_array_equal, assert_array_almost_equal
7	7	import netCDF4

58	58	self.assertTrue(bar.shape == (n1dim,n2dim,n3dim))
59	59	# check data.
60	60	#assert_array_almost_equal(bar[:,:,:], ranarr)
61		assert_array_almost_equal(bar[:,:,:], 2.*NP.ones((n1dim,n2dim,n3dim),ranarr.dtype))
	61	assert_array_almost_equal(bar[:,:,:], 2.*np.ones((n1dim,n2dim,n3dim),ranarr.dtype))
62	62	f.close()
63	63
64	64	if __name__ == '__main__':

-3

test/tst_vars.py less more

1	1	import unittest
2	2	import os
3	3	import tempfile
4		import numpy as NP
	4	import numpy as np
5	5	from numpy.random.mtrand import uniform
6	6	from numpy.testing import assert_array_equal, assert_array_almost_equal
7	7	import netCDF4

66	66	assert v1.size == DIM1_LEN * DIM2_LEN * DIM3_LEN
67	67	assert len(v1) == DIM1_LEN
68	68
69		#assert NP.allclose(v1[:],randomdata)
	69	#assert np.allclose(v1[:],randomdata)
70	70	assert_array_almost_equal(v1[:],randomdata)
71	71	# check variables in sub group.
72	72	g = f.groups[GROUP_NAME]

85	85	assert v2.dimensions == (DIM2_NAME,DIM3_NAME)
86	86	assert v1.shape == (DIM1_LEN,DIM2_LEN,DIM3_LEN)
87	87	assert v2.shape == (DIM2_LEN,DIM3_LEN)
88		#assert NP.allclose(v1[:],randomdata)
	88	#assert np.allclose(v1[:],randomdata)
89	89	assert_array_almost_equal(v1[:],randomdata)
90	90	f.close()
91	91