BUG: clip where the bound is a series with NA values returns NA · Issue #40420 · pandas-dev/pandas (original) (raw)

In [1]: import pandas as pd

In [2]: pd.__version__
Out[2]: '1.2.3'

In [3]: tmp = pd.DataFrame({'f': [1,2]})

In [4]: import numpy as np

In [5]: tmp['h'] = np.nan
#This is the expected behaviour
In [6]: tmp['f'].clip(0, np.inf)
Out[6]: 
0    1
1    2
Name: f, dtype: int64
# This seems wrong:
In [7]: tmp['f'].clip(0, tmp['h'])
Out[7]: 
0   NaN
1   NaN
Name: f, dtype: float64