CODEWITHSUNDEEP
pythonpandas
Gary
Gary

Reputation: 899

group by one column, find max of another and display the remaining columns

I have a dataframe df. I need to group by column a, find max of column b but want to display all the columns in df as well.

For example -

df:

    a       b       c      d
0   259992  20     55      abc
1   259992  10     12      def
2   1912687 11     11      ghi

My code to find max of b:

df.groupby('a')['b'].max().reset_index()

This gives me a partial result:

    a       b   
0   259992  20 
1   1912687 11

What I want to obtain is:

    a       b       c      d
0   259992  20     55      abc
1   1912687 11     11      ghi

Any tips on how to do this in the most efficient manner ?

Upvotes: 1

Views: 39

Answers (1)

anky
anky

Reputation: 75080

Looks like you need idxmax:

df.loc[df.groupby('a')['b'].idxmax()]

         a   b   c    d
0   259992  20  55  abc
2  1912687  11  11  ghi

Upvotes: 2

Related Questions