CODEWITHSUNDEEP
pandaspandas-groupbyrank
CutePoison
CutePoison

Reputation: 5355

Ranking order within groupby groups

Say I have the following dataframe df_grouped:


                     Sale
custid | product   |
-------+-----------+------
1      |   iPhone  | 3000
       |   android | 2000  
2      |   iPhone  | 1500
       |   android | 2500

and I want a ranking within each group (customerid) ending with


                     Sale
custid | product   |
-------+-----------+------
1      |   iPhone  | 1
       |   android | 2
2      |   iPhone  | 2
       |   android | 1

If I just do df_grouped.rank() I get a global ranking


                     Sale
custid | product   |
-------+-----------+------
1      |   iPhone  | 1
       |   android | 3
2      |   iPhone  | 4
       |   android | 2

but I want the ranking within each group. I can loop over each custid and apply rank to each of those groups, but isn't there a faster/cleaner way?

Upvotes: 1

Views: 49

Answers (1)

jezrael
jezrael

Reputation: 862511

Use GroupBy.rank:

df['Rank'] = df.groupby(level='custid')['Sale'].rank()

Upvotes: 1

Related Questions