pandas dataframe how to take only rows with max value on one col, per group

Asked Feb 14 '22 at 14:36

Active Feb 14 '22 at 14:36

Viewed 17 times

I have a dataframe with categorical value, numerical values and a counter. I want to keep only rows with the highest counter per category. So, if my dataframe is:

df = Tr. N1. N2. counter
     T1.  0.  3.    0 
     T1.  2.  3.    0
     T1.  7.  1.    1 
     T2.  2.  7.    0
     T2.  0.  9.    1 
     T2.  5.  6.    2
     T2.  3.  2.    2 
     T3.  7.  8.    0

For each value of Tr, I want to keep only rows with the highest counter. So, the new df will be:

df = Tr. N1. N2. counter

     T1.  7.  1.    1  
     T2.  5.  6.    2
     T2.  3.  2.    2 
     T3.  7.  8.    0

What is the best way to do so?

asked Feb 14 '22 at 14:36

Cranjis

1,590
8
31
64

What happens when there are multiple rows for the same `Tr` with highest `counter` value? Pick random or is there a logic there too? – Mayank Porwal Feb 14 '22 at 14:40
@MayankPorwal to take all of them (see rows of T2) – Cranjis Feb 14 '22 at 14:40
2

Use: `df[df.groupby('Tr')['counter'].transform('max').eq(df['counter'])]` – Mayank Porwal Feb 14 '22 at 14:42

pandas dataframe how to take only rows with max value on one col, per group

0 Answers0