ncdata.netcdf4 module#

Interface routines for converting data between ncdata and netCDF4.

Converts ncdata.NcData to and from netCDF4.Dataset objects.

ncdata.netcdf4.from_nc4(nc4_dataset_or_file, dim_chunks=None)#

Load NcData from a netCDF4.Dataset or netCDF file.

Parameters:

  • nc4_dataset_or_file (Dataset | Group | Path | str) – source of load data. Can be either a netCDF4.Dataset, a netCDF4.Group, a pathlib.Path or a string.

  • dim_chunks (Dict[str, int | str]) – a dictionary of chunk sizes (number, or -1 or “auto”) for specific dimensions, specified by dimension name. Defaults to “auto” for all unspecified dimensions.

Returns:

ncdata

Return type:

NcData

Examples

For example, to avoid cases where a simple dask from_array(chunks="auto") will fail

>>> from ncdata.netcdf4 import from_nc4
>>> ds = from_nc4(testfile_path, dim_chunks={"x": 15})
>>> ds.variables["vx"].data.chunksize
(15,)
>>>

See also : Load a file containing variable-width string variables : This illustrates a particular case which does encounter an error with dask “auto” chunking, and therefore also fails with a plain “from_nc4” call. The dim_chunks keyword enables you to work around the problem.

ncdata.netcdf4.to_nc4(ncdata, nc4_dataset_or_file, var_kwargs=None)#

Save NcData to a netCDF file.

Parameters:

  • ncdata (NcData) – input data

  • nc4_dataset_or_file (netCDF4.Dataset or pathlib.Path or str) – output filepath or netCDF4.Dataset to write into

  • var_kwargs (dict or None) –

    additional keys for variable creation. If present, each entry var_kwargs[<var_name>] is itself a dictionary, containing entries of the form <kwarg_name>: <kwarg_value>, which specify additional keyword controls passed to netCDF4.Dataset.createVariable() for this variable. This allows control of e.g. compression or chunking.

    Keyword controls for variables in a sub-group can be given in a var_kwargs['/<group-name>'] entry, which is itself a var_kwargs - like dictionary.

    Note: this must not include keywords controlled by the to_nc4 operation itself. That is, any of : [“data”, “dimensions”, “datatype”, “fill_value”].

Return type:

None

Notes

If a filepath is provided, a file is written and closed afterwards. If a dataset is provided, it must be writeable and remains open afterward.

On this page