Pandas groupBy two columns and keeping the columns

Question

I have a dataframe that looks like:

param1   param2   num1    num2
A          B        4      0.3
A          C        5      0.1
A          B        6      0.5
B          F        3      0.7
C          A        1      0.8
C          A        2      1.1

and I want something like

param1   param2   num1    num2
A          B        10     0.8
A          C        5      0.1
B          F        3      0.7
C          A        3      1.9

If I use something like: df.groupby(['param1', 'param2']).sum() I end up with the columns 'param1' and 'param2' fused as an index, but I don't want that. I just need them to stay as they are, but if I have some lines with the same param1 and param2, sum the 'nums' associated to them.

Use `df.groupby(['param1', 'param2'], as_index=False).sum()` or `df.groupby(['param1', 'param2']).sum().reset_index()` — jezrael, Sep 03 '21 at 09:00

mozway · Accepted Answer · 2021-09-03T09:01:10.360

3

Use the as_index=False parameter:

df.groupby(['param1', 'param2'], as_index=False).sum()`

edited Sep 03 '21 at 09:01

answered Sep 03 '21 at 09:00

mozway

194,879
13
39
75

2

Why answering for dupes? – jezrael Sep 03 '21 at 09:00
because it takes me much more time to search for the potential duplicates, and that in fine, it will help OP better to just provide the answer – mozway Sep 03 '21 at 09:02
yes, but then it means need points, because not change to wiki, right? – jezrael Sep 03 '21 at 09:02
I didn't get what you mean – mozway Sep 03 '21 at 09:03
Sure, do it if you want – mozway Sep 03 '21 at 09:04
1

ya, its happens answeing for dupes, but why need points for it? If dont need points wiki is best use here in my opinion. There is answer and not get points for 1000 times answered questions in dupes. – jezrael Sep 03 '21 at 09:05
sorry, i imagined this question was already asked, but I couldn't find the answer; I probably didn't use the proper search terms – Reine Baudache Sep 03 '21 at 09:26
hmmm, so no wiki? – jezrael Sep 03 '21 at 09:33
1

@jezrael i thought you did it, I told you you could. I'm also busy IRL ;) – mozway Sep 03 '21 at 09:47
@mozway - I cannot do it ;) – jezrael Sep 03 '21 at 09:47

Pandas groupBy two columns and keeping the columns

1 Answers1