replace inf values with 0 in netcdf in python

Question

I can replace nan values in NetCDF using xarray like this:

hndl_nc = hndl_nc.fillna(0.0)

Is there a way to replace inf values with 0 as well?

Answer 1

Numpy's isfinite captures both np.inf and np.nan.

hndl_nc.where(hndl_nc.apply(np.isfinite)).fillna(0.0)

Note that the next version of xarray (0.10) will allow you to specify the fill value in the where method directly.

Answer 2

Try something like:

def replace_inf_with_zero(dataset):
    for varname in dataset.data_vars.keys():
        array = hndl_nc[varname]
        array[array == np.inf] = 0.