CODEWITHSUNDEEP
pythonpandasdataframe
ah bon
ah bon

Reputation: 10051

Groupby multiple columns count size and calculate mean of another column in Pandas

I have a dataframe as follows:

   index    col1    col2   col3  col4                           col5
0      0  Week_1   James   John     1                  when and why?
1      1  Week_1   James   John     3             when and why? How?
2      2  Week_2   James   John     2  How far is it? Are you going?
3      3  Week_2    Mark    Jim     3              Do you know when?
4      4  Week_2  Andrew  Simon     3                     What time?
5      5  Week_2  Andrew  Simon     6                     What time?

How could I groupby col2 and col3 then calculate mean and count numbers of col2 and col3?

df.groupby(['col2','col3'], as_index=False).agg({'col4':'mean'}).reset_index()

Output:

   index    col2   col3  col4
0      0  Andrew  Simon   4.5
1      1   James   John   2.0
2      2    Mark    Jim   3.0

df.groupby(['col2','col3']).size().reset_index()

Output:

     col2   col3  0
0  Andrew  Simon  2
1   James   John  3
2    Mark    Jim  1

How could I get result like this? Thanks.

   index    col2   col3  mean  count
0      0   James   John   2.0      3
1      3    Mark    Jim   3.0      1
2      4  Andrew  Simon   4.5      2

Upvotes: 0

Views: 667

Answers (1)

Allen Qin
Allen Qin

Reputation: 19957

You can use groupby and agg(may need to use pandas 0.25+).

(
    df.groupby(['col2','col3'])
    .agg(index=('index', 'first'),
         mean=('col4', 'mean'),
         count=('col4', 'size'))
    .reset_index()
    .sort_values(by='index')
)

    col2    col3    index   mean    count
1   James   John    0       2.0     3
2   Mark    Jim     3       3.0     1
0   Andrew  Simon   4       4.5     2

Upvotes: 3

Related Questions