Pandas groupby each group and divide each row by its sum

Asked Jan 29 '21 at 02:51

Active Jan 29 '21 at 03:33

Viewed 2,232 times

I have a df

df = pd.DataFrame({'group':['a','a','b','b','c','c','c'], 'score':[1,2,3,4,5,6,7]})

group score
0   a   1
1   a   2
2   b   3
3   b   4
4   c   5
5   c   6
6   c   7

I want to divide each row by the sum of its group. So for group a, each of the 2 rows will divide by 3 (2+1), for group c, each of the 3 rows will divide by 18 (5+6+7), giving:

group score
0   a   0.33
1   a   0.67
2   b   0.43
3   b   0.57
4   c   0.28
5   c   0.33
6   c   0.39

I was thinking of using a df.groupby('group')['score'].apply(lambda x: x / sum???) approach but can't figure it out. Any help will be appreciated!

asked Jan 29 '21 at 02:51

TYL

1,577
20
33

4

`df['score']/df.groupby('group')['score'].transform('sum')` – Quang Hoang Jan 29 '21 at 02:54
@QuangHoang do post as answer since I was about to post the same but u already commented. thanks – Akshay Sehgal Jan 29 '21 at 02:55
2

@AkshaySehgal This surely has a duplicate somewhere on SO, so I wouldn't post an answer :-). – Quang Hoang Jan 29 '21 at 02:56
@QuangHoang Yeah I thought there might be a solution somewhere on SO, but can't seem to find it! – TYL Jan 29 '21 at 02:57
It's surprising that the recommended posts that show up when marking something duplicate are not that great and instead googling the problem shows the top SO posts directly :) – Akshay Sehgal Jan 29 '21 at 03:03
@AkshaySehgal You're correct, SO search is terrible. I never use it. – Trenton McKinney Jan 29 '21 at 03:38

Pandas groupby each group and divide each row by its sum

0 Answers0