How to keep duplicates in a pandas groupby max?

Asked Apr 26 '20 at 12:55

Active Apr 26 '20 at 12:55

Viewed 15 times

I have a dataframe df

        date  close
0 2006-01-02  143.2
1 2006-01-07  165.2
2 2008-09-19  198.0
3 2008-12-20  198.0

I want to groupby year and find the maximum close such that duplicates are kept.

So far what I have is

df.groupby(df.date.dt.year).max()[["close"]]
           date  close
date                  
2006 2006-01-07  165.2
2008 2008-12-20  198.0

whereas what I want is the two close values in 2008 since they are identical

           date  close
date                  
2006 2006-01-07  165.2
2008 2008-09-19  198.0
2008 2008-12-20  198.0

How do I do this?

asked Apr 26 '20 at 12:55

PyRsquared

1

Check solution with `groupby.transform` – jezrael Apr 26 '20 at 12:58
```df.groupby(df.date.dt.year).filter(lambda x: x.close.eq(x.close.max()).all())``` maybe? this will be slower than @jezrael suggestion of transform – sammywemmy Apr 26 '20 at 12:58
1

Use `df[df.groupby(df.date.dt.year)['close'].transform('max').eq(df["close"])]` – jezrael Apr 26 '20 at 13:00

0 Answers0