Issue

This Content is from Stack Overflow. Question asked by Patrik

I have a dataset in this form:

col_1     col_2    col_3    col_4     col_5

0          0        0        0         Nan
0          1        Nan      1         1
1          0        1        0         Nan
0          0        0        0         0

Now, I want to create a new column such that for any particular row, if any of the column values is 1, then the output should be 1.. For example, in above dataset, the new column will have second and third row as 1…

SO, I have tried this approach:

if ((df['col_1]==1) | (df['col_2']==1) | (df['col_3']==1) | (df['col_4']==1) | (df['col_5']==1)):
    df['new_column']=1
else:
    df['new_column']=0

This code was giving me error.

So, I tried a different approach.

lists = ['col_1','col_2','col_3','col_4','col_5']
for i in lists:
   if(df[i]==1):
       df['new_column]==1
    else:
        df['new_column']==0

This code is again giving me wrong values…

Can someone please help me in solving it as I am beginner in pandas and stuck in this problem.

Solution

Use comparison to 1 (with eq), aggregate with any per row (axis=1), then perform boolean indexing:

df.loc[df.eq(1).any(axis=1), 'new_col'] = 1

output:

   col_1  col_2 col_3  col_4 col_5  new_col
0      0      0     0      0   Nan      NaN
1      0      1   Nan      1     1      1.0
2      1      0     1      0   Nan      1.0
3      0      0     0      0     0      NaN

If you prefer a 0/1 output:

df['new'] = df.eq(1).any(1).astype(int)

output:

   col_1  col_2 col_3  col_4 col_5  new
0      0      0     0      0   Nan    0
1      0      1   Nan      1     1    1
2      1      0     1      0   Nan    1
3      0      0     0      0     0    0

This Question was asked in StackOverflow by Patrik and Answered by mozway It is licensed under the terms of CC BY-SA 2.5. - CC BY-SA 3.0. - CC BY-SA 4.0.