How to add up pandas dfs that contains some nan value while retaining all values (i.e. 0+nan=0)?

Question

As title, how can I add up pandas dfs that contains some nan value while retaining all values in those entries(i.e. 0+nan=0)?

DF1=

A1	B1	C1	D1
0	0	nan	nan
0	1	1	nan

DF2=

A2	B2	C2	D2
nan	nan	nan	nan
nan	1	1	1

Expected output=

A sum	B sum	C sum	D sum
0	0	nan	nan
0	2	2	1

Sorry if the question is too basic, thanks for answering.

Answer 1

You need to use add with fill_value=0 and to avoid index alignment by covering to_numpy:

df1.add(df2.to_numpy(), fill_value=0)

Output:

    A1   B1   C1   D1
0  0.0  0.0  NaN  NaN
1  0.0  2.0  2.0  1.0

With column names:

(df1.add(df2.to_numpy(), fill_value=0)
    .rename(columns=lambda c: c[:-1] + ' sum')
)

Output:

   A sum  B sum  C sum  D sum
0    0.0    0.0    NaN    NaN
1    0.0    2.0    2.0    1.0