NoData values in SPI output

[mortezaomidi]

Thank you for your valuable work.
I ran the SPI function for a NetCDF file (time series from 2000, 2023), but unfortunately many pixels calculated as NoData.

Could this line of code be the problem?
ds_In = ds_In.where(np.isinf(ds_In) == False) #= np.nan #Change infinity to NaN

Do you have an idea to solve the problem?