Creating new backends#

Introduction#

You can read different type of files) in xr.open_dataset by specifying the engine to be used:

import xarray as xr
xr.open_dataset("my_file.grib" , engine="cfgrib")

For each available engine there is an underlying backend, that reads the data and pack them in a dataset.

Xarray bundles several backends internally for the following formats:

  • netcdf4 - netCDF4

  • scipy - netCDF3

  • zarr - Zarr

  • pydap - DAP

  • …

External Backends that use the new backend API (xarray >= v0.18.0) that allows to add support for backend without any change to Xarray

Why using the Xarray backend API#

  • Your users don’t need to learn a new interface that is they can use xr.open_dataset with engine kwarg.

  • With little extra effort you can have lazy loading with Dask. you have to implement a function for reading blocks and Xarray will manage lazy loading with Dask for you

  • It’s easy to implement: you don’t need to integrate any code in Xarray

Next#

See the documentation for more.

Follow the tutorial on creating a new backend for binary files.