How do I …

How do I…


add a DataArray to my dataset as a new variable

my_dataset[varname] = my_dataArray or Dataset.assign() (see also Dictionary like methods)

add variables from other datasets to my dataset


add a new dimension and/or coordinate

DataArray.expand_dims(), Dataset.expand_dims()

add a new coordinate variable


change a data variable to a coordinate variable


change the order of dimensions

DataArray.transpose(), Dataset.transpose()

reshape dimensions

DataArray.stack(), Dataset.stack(), Dataset.coarsen.construct(), DataArray.coarsen.construct()

remove a variable from my object

Dataset.drop_vars(), DataArray.drop_vars()

remove dimensions of length 1 or 0

DataArray.squeeze(), Dataset.squeeze()

remove all variables with a particular dimension


convert non-dimension coordinates to data variables or remove them

DataArray.reset_coords(), Dataset.reset_coords()

rename a variable, dimension or coordinate

Dataset.rename(), DataArray.rename(), Dataset.rename_vars(), Dataset.rename_dims(),

convert a DataArray to Dataset or vice versa

DataArray.to_dataset(), Dataset.to_array(), Dataset.to_stacked_array(), DataArray.to_unstacked_dataset()

extract variables that have certain attributes


extract the underlying array (e.g. numpy or Dask arrays)

convert to and extract the underlying numpy array


find out if my xarray object is wrapping a Dask Array


know how much memory my object requires

DataArray.nbytes, Dataset.nbytes

Get axis number for a dimension


convert a possibly irregularly sampled timeseries to a regularly sampled timeseries

DataArray.resample(), Dataset.resample() (see Resampling and grouped operations for more)

apply a function on all data variables in a Dataset

write xarray objects with complex values to a netCDF file

Dataset.to_netcdf(), DataArray.to_netcdf() specifying engine="h5netcdf", invalid_netcdf=True

make xarray objects look like other xarray objects

ones_like(), zeros_like(), full_like(), Dataset.reindex_like(), Dataset.interp_like(), Dataset.broadcast_like(), DataArray.reindex_like(), DataArray.interp_like(), DataArray.broadcast_like()

Make sure my datasets have values at the same coordinate locations

xr.align(dataset_1, dataset_2, join="exact")

replace NaNs with other values

Dataset.fillna(), Dataset.ffill(), Dataset.bfill(), Dataset.interpolate_na(), DataArray.fillna(), DataArray.ffill(), DataArray.bfill(), DataArray.interpolate_na()

extract the year, month, day or similar from a DataArray of time values

obj.dt.month for example where obj is a DataArray containing datetime64 or cftime values. See Datetime components for more.

round off time values to a specified frequency

obj.dt.ceil, obj.dt.floor, obj.dt.round. See Datetime components for more.

make a mask that is True where an object contains any of the values in a array

Dataset.isin(), DataArray.isin()

Index using a boolean mask

Dataset.query(), DataArray.query(), Dataset.where(), DataArray.where()

preserve attrs during (most) xarray operations