How to get max of counts for groupby

I have a dataframe. I want to group by rows on some columns and then form a count column and then get the max of counts and create a column for it and attach it to dataframe.

I tried:

    df["max_pred"] = df.groupby(['fid','prefix','pred_text1'], 
                             sort=False)["pred_text1"].transform("max")

However it lists a row with max repeat for pred_text1, but I want the number of reparation for it

For example:

A  B  C
a  d  b
a  d  b
a  d  b
a  d  a
a  d  a
b  b  c
b  b  c
b  b  d

If I group by A and B and then count of C and get max and store that in new column F, I expect:

 A  B  F
a  d  3
a  d  3
a  d  3
a  d  3
a  d  3
b  b  2
b  b  2
b  b  2

>Solution :

You can use groupby.transform with value_counts:

df['F'] = (df.groupby(['A', 'B'])['C']
             .transform(lambda g: g.value_counts(sort=False).max())
          )

Variant with collections.Counter:

from collections import Counter

df['F'] = (df.groupby(['A', 'B'])['C']
             .transform(lambda g: max(Counter(g).values()))
          )

Output:

   A  B  C  F
0  a  d  b  3
1  a  d  b  3
2  a  d  b  3
3  a  d  a  3
4  a  d  a  3
5  b  b  c  2
6  b  b  c  2
7  b  b  d  2