Pandas将选定的列合并为1

Question

I have a df like this:

ID1 ID2 Day Text1 Text2 Text3 ....
111 A   1   a     b     c
222 B   2   i     j     k
333 C   3   x     y     z

My goal is to create a new columns that contain all values of Text1, Text2, Text3, and so on.

ID1 ID2 Day Text1 Text2 Text3 ....  Text
111 A   1   a     b     c           a, b, c...
222 B   2                          
333 C   3   x           y           x, y, .... 

I've tried:

list(zip(df.Text1,df.Text2,df.Text3,...)): 

This works but the format isn't desirable.

And:

df.apply(lambda x: ', '.join(x.astype(str)), axis=1): 

This gives the desired format but the answer will contains all fields.

What would be the best approach this? Many thanks!

Answer 1

Vectorized solution:

In [65]: df['Text'] = df.filter(regex='^Text\d+').add(', ').sum(1).str.rstrip(', ')

In [66]: df
Out[66]:
   ID1 ID2  Day Text1 Text2 Text3     Text
0  111   A    1     a     b     c  a, b, c
1  222   B    2     i     j     k  i, j, k
2  333   C    3     x     y     z  x, y, z

Answer 2

Your code is very close. You just need to use apply on a df[text_cols] where text_cols is a list of the columns you want to merge into a new one.

df['Text'] = df[text_cols].apply(lambda x: ''.join(x), axis=1)

Answer 3

There is also a vectorized join :

>>> df['Text'] = df.filter(regex='^Text\d+').sum(1).str.join(', ')
>>> df
   ID1 ID2  Day Text1 Text2 Text3     Text
0  111   A    1     a     b     c  a, b, c
1  222   B    2     i     j     k  i, j, k
2  333   C    3     x     y     z  x, y, z

Answer 4

The other solutions are awesome, and I would like offer an answer that uses the cat() function.

df['text'] = df[0].str.cat([df[i] for i in df.columns[1:]],sep=',')

Hope it helps : )