CODEWITHSUNDEEP
pythonpandasdataframe
Soufiane Sabiri
Soufiane Sabiri

Reputation: 768

Join groupby column with a comma in a Pandas DataFrame

I have a dataset like this:

>>> df = pd.DataFrame({'id_sin':['s123','s123','s124','s124'],
                       'raison':['first problem','second problem','album','dog']
    })
>>> df
  id_sin          raison
0   s123   first problem
1   s123  second problem
2   s124           album
3   s124             dog

This is the expected output:

  id_sin                         raison
0   s123  first problem, second problem
1   s124                     album, dog

What I tried:

df['raison'] = df.groupby('id_sin')['raison'].apply(lambda x: ', '.join(x))

But doesn't work... what am I missing? Thanks for help!

Upvotes: 1

Views: 1445

Answers (2)

Scott Boston
Scott Boston

Reputation: 153500

Try using agg:

df.groupby('id_sin')['raison'].agg(', '.join).reset_index()

Output:

  id_sin                         raison
0   s123  first problem, second problem
1   s124                     album, dog

Upvotes: 3

Quang Hoang
Quang Hoang

Reputation: 150785

Try changing the groups to lists:

df.groupby(['id_sin']).raison.apply(lambda x: ', '.join(list(x)))

After testing your code, it turns out that you should not do df['raison'] =... because df.groupby('id_sin')['raison'].apply(lambda x: ', '.join(x)) has length 2 with different index than df, which has length 4.

Upvotes: 1

Related Questions