Replace certain values in pandas dataframe with mode of that row

Question

I have a DataFrame like this:

ColA      ColB  ColC  ColD  ColE

Male       1     0     ?     1
Female     1     1     ?     1
Male       0     0     0     1
Female     1     0     0     ?
Female     1     1     1     ?

It has a number of ? throughout. I'd like to replace these with the mode (most common value of each row) - excluding the first column.

I've done something like this, but doesn't return what I want:

df.replace("?", np.nan, inplace = True)     
df_new = df.apply(lambda row: row.fillna(row[1:].mode()), axis = 1)

Answer 1

For fillna we can only fillna with column with series, so we need T , and for the mode we can do outside the apply

df_new = df.T.fillna(df.iloc[:,1:].astype(float).mode(axis=1)[0]).T
df_new
     ColA ColB ColC ColD ColE
0    Male    1    0  1.0    1
1  Female    1    1  1.0    1
2    Male    0    0    0    1
3  Female    1    0    0  0.0
4  Female    1    1    1  1.0