Pandas replace by condition in the 1st row

Question

I have a table:

Object	Col1	Col2	Col3	Col4
reference	10	0	7	0
Obj1	14	9	1	30
Obj2	1	16	0	0
Obj3	9	21	3	17
Obj4	11	0	4	22

And I want to transform it by condition: if 1st row of any column is =0, then other rows of this column can be either 0 or 1. If the current value is >0, than it becomes 1, if it's 0, than it remains 0.

Resulting table is:

Object	Col1	Col2	Col3	Col4
reference	10	0	7	0
Obj1	14	1	1	1
Obj2	1	1	0	0
Obj3	9	1	3	1
Obj4	11	0	4	1

I've tried this variant:

for column in df:
    df[column] = np.where((df[column][0] == 0) & (df[column] != 0), 1, 0)

but it resets to zero all columns, where 1st row isn't equal to 0. How can I do this right?

Answer 1

You can select first row by df.iloc[0] and then check which columns in this row are equal 0. Then change the particular columns:

first_row = df.iloc[0]
zero_columns = first_row[first_row == 0].index

df.loc[:, zero_columns] = (df.loc[:, zero_columns] > 0).astype(int)
print(df)

Prints:

      Object  Col1  Col2  Col3  Col4
0  reference    10     0     7     0
1       Obj1    14     1     1     1
2       Obj2     1     1     0     0
3       Obj3     9     1     3     1
4       Obj4    11     0     4     1