set value of column1 to value of column2 based on value of column1

Question

Trying to change values in column 1 to the values found in column 2 if the values in column 1 == 'nan'.

for row in df.column_1:
    if row == 'nan':
       df.column_1 == df.column_2

But no changes are seen.

Desired Before and After: [sorry, graphic won't upload at the moment]

I should have mentioned that I tried that too. Does not work. — spacedustpi, Aug 07 '18 at 14:47

BENY · Accepted Answer · 2018-08-07T15:21:27.723

1

1st 'nan' is not NaN, if we want to use fillna we need replace

df.column_1=df.column_1.replace('nan',np.nan).fillna(df.column_2)

or we just using np.where

df.column_1=np.where(df.column_1=='nan',df.column_2,df.column_1)

edited Aug 07 '18 at 15:21

answered Aug 07 '18 at 14:51

BENY

Thanks Wen, the second solution (np.where) worked. The first one (.replace) did not. – spacedustpi Aug 07 '18 at 15:07
Taking out the curly brackets allows the first solution to work for my issue. I tried to take out the curlies in an edit, but it said I need to change at least 6 characters to make an edit. Strange. – spacedustpi Aug 07 '18 at 15:20
1

@spacedustpi yep , I miss type the dataframe replace to series sorry about that – BENY Aug 07 '18 at 15:21

1 Answers1