基于值的列的Pandas数据框组合

Question

I have a pandas dataframe with 13 columns - ID(unique identifier),A1,A2,..A12. all A columns can have 2 values- 0 or 1

 d = {'ID': ['ID1', 'ID2','ID3', 'ID4'], 'A1': [0,0,0,1], 'A2': [1,0,0,1], 'A3': [0,0,0,0], 'A4': [1,1,0,1], 'A5': [0,0,0,1]
    , 'A6': [0,1,0,0], 'A7': [1,1,0,1], 'A8': [1,0,0,0], 'A9': [1,1,0,1], 'A10': [0,1,0,0], 'A11': [1,1,1,0], 'A12': [1,0,1,1]}
df = pd.DataFrame(data=d)
df

I want to add a new column, A_combined where its value is a combination of the 12 other columns, if their value is 1. For example, if the row is

ID1 1 0 0 0 0 1 0 0 1 0 1 0 

then A_combined will have the value A1_A6_A9_A11

Any help would be highly appreciated!

UPDATE

I am able to achieve a restructring of the dataframe, using @wen 's suggestions:

import numpy as np

v=df.iloc[:,:12]

test=v.mul(v).replace(0,np.nan).stack().reset_index()

test

Here 'test' has column names at row level. Any suggestions on next steps to combine row values by index? Thanks!

Answer 1

v=dd.iloc[:,1:]
dd['Acombine']=v.mul(v.columns).replace('',np.nan).stack().groupby(level=0).apply('_'.join)
dd
Out[859]: 
    ID  A1  A2  A3  A12 Acombine
0  ID1   0   0   1    1   A3_A12

Answer 2

Not to sure if I am following your example completely (ie "combination of the 12 other columns, if their value is 1", if what is 1, the first column?).

df.loc[df['A1'] == 1, 'A_'] = [df['A1'].astype(str)+df['A2'].astype(str)+df['A3'].astype(str)]

This code reads like so: if column 'A1' is equal to 1, then create and fill column 'A_', with the values from columns A1, A2, and A3. The far right part of the code could be modified to include all 12 columns.

Answer 3

I believe the answer below is what you're looking for without having to restructure the data. It uses a temporary dataframe which has the instances of 1 replaced with the colum header you want. Then it creates a new column with the values in each row joined as you want, added back to the original dataframe.

df2 = pd.DataFrame()
for col in df.columns:
    df2[col] = df[col].replace(1, col)

def func(x):
    return '_'.join(str(i) for i in x if i != 0)

df.assign(combined = df2.apply(func, axis=1))

    A1  A10 A11 A12 A2  A3  A4  A5  A6  A7  A8  A9  combined
0   0   0   1   1   1   0   1   0   0   1   1   1   A11_A12_A2_A4_A7_A8_A9
1   0   1   1   0   0   0   1   0   1   1   0   1   A10_A11_A4_A6_A7_A9
2   0   0   1   1   0   0   0   0   0   0   0   0   A11_A12
3   1   0   0   1   1   0   1   1   0   1   0   1   A1_A12_A2_A4_A5_A7_A9