Pandas dataframe groupby sum strings in reverse order-CodePudding

I have a DataFrame like this:

colA    colB 
1       aaa
1       rrr
1       www
2       bbb
2       ccc
2       sss
...

I would like to convert the DataFrame as follows

colA    Sum
1       wwwrrraaa
2       ssscccbbb
...

I tried

df.groupby(['colA'])['colB'].sum().reset_index()

but the sum of strings are reversed. Is there an elegant way to do this?

CodePudding user response：

Don't use sum to concatenate strings. It looks fancy but it's quadratic and should be considered bad practice. In python is use function join

df = df[::-1].groupby('colA')['colB'].agg(''.join).reset_index()

CodePudding user response：

Reverse the DataFrame; then groupby sum:

out = df[::-1].groupby('colA', as_index=False)['colB'].sum()

Output:

   colA       colB
0     1  wwwrrraaa
1     2  ssscccbbb