Pandas - Fill Column Values based on conditions on Two other columns

Asked Sep 05 '19 at 11:30

Active Sep 05 '19 at 11:40

Viewed 25 times

I want to fill the values of column 3 based on condition on column 1 and 2 without having to import numpy and use np.where

DataFrame:

1  CNT T
2  T   T
3  T   NT
4  CNT INT

Condition: if value in Column1 != 'T' and value in Column2 == 'T' then value in Column3 = value in Column1 (='CNT') else value in Column3 == value in Column2

Output:

1  CNT  T   CNT
2  T    T   T
3  T    NT  NT
4  CNT  INT INT

What I have tried is:

Dataframe[3] = Dataframe.loc[(Dataframe[1] != 'T') & (Dataframe[2] == 'T'), 1]

but that creates Output:

1  CNT  T   CNT
2  T    T   NaN
3  T    NT  NaN
4  CNT  INT NaN

edited Sep 05 '19 at 11:40

asked Sep 05 '19 at 11:30

bloo

Have you tried `np.where()` ? – samkart Sep 05 '19 at 11:32
You are close, need `Dataframe[3] = np.where((Dataframe[1] != 'T') & (Dataframe[2] == 'T'), Dataframe[1], Dataframe[2])` – jezrael Sep 05 '19 at 11:32
Or `df.loc[df[1].ne('T') & df[2].eq('T'), 3] = df[1]` – yatu Sep 05 '19 at 11:33
apologies if this is a duplicate, @yatu that returns the same as mine. samkart and jezrael: is there a way without having to import numpy? – bloo Sep 05 '19 at 11:38
2

@bloo - use `Dataframe[3] = pd.np.where((Dataframe[1] != 'T') & (Dataframe[2] == 'T'), Dataframe[1], Dataframe[2])` – jezrael Sep 05 '19 at 11:52
awesome! didn't know that could be done! thanks! – bloo Sep 05 '19 at 11:57

Pandas - Fill Column Values based on conditions on Two other columns

0 Answers0