how can i sum a pandas dataframe for every combination?

Question

I have a dataframe that looks like:

user   letter   time
user1    a       2
user2    b       5
user1    b       3
user1    a       5

I try to get a dataframe like this, where time is summed for the same letter for every user, but can't find the solution for this problem:

user   letter   time
user1    a       7
user2    b       5
user1    b       3

score 0 · Answer 1 · answered May 20 '19 at 08:32

Use df.groupby(['user', 'letter'])['time'].sum().

Output:

user   letter
user1  a         7
       b         3
user2  b         5

If you want to keep the original columns, you can pass as_index=False to groupby:

df.groupby(['user', 'letter'], as_index=False)['time'].sum()

Output:

    user letter  time
0  user1      a     7
1  user1      b     3
2  user2      b     5

score 0 · Answer 2 · answered May 20 '19 at 08:32

Try this groupby:

print(df.groupby(['user', 'letter']).sum())

Output:

              time
user  letter      
user1 a          7
      b          3
user2 b          5

A little nicer one:

print(df.groupby(['user', 'letter'], as_index=False).sum())

Output:

    user letter  time
0  user1      a     7
1  user1      b     3
2  user2      b     5

2 Answers2