I checked docs of pandas isna and numpy isnan.
They seemed to say thaat numpy.isnan detects NaN only and pandas.isna detects both NaN and None type.
Is that so?
What's the difference between pandas.isna and numpy.isnan?
Viewed 1872
1 Answers
Yes, this is correct.
np.isnan checks if your value is np.nan or your iterable (array,list) contains np.nan.
For example (from their documentation):
np.isnan([np.log(-1.),1.,np.log(0)]) results array([ True, False, False]) this is because np.log(-1) is not defined and results np.nan. np.isnan does not detects python None.
pandas.isna on the other hand lays above numpy and detects both np.nan and None values.
Related