Choosing 3 (or n) observation from given index

Question

I'm trying to take the 3 highest observations for each index. For instance, I have

census=pd.Series([2000,4432,5435,43252,63463,423432,3525,54353,6363]) 
census.index=['AL','AL','AL','AL','AK','AK','AK','AK','AK']

I want to get 3 highest observation for AL and AK and get it as a different series. Thanks.

Start with `s = census.groupby(level=0).nlargest(3)` and then figure out how to fix the result yourself. — cs95, Dec 06 '18 at 03:38
Thanks, that answers my question, but I don't know how to mark this as the correct answer. Appreciate it. — Sandeep, Dec 06 '18 at 03:58
At least don't accept the other answer until they fix their code. — cs95, Dec 06 '18 at 03:59

Axel Puig · Answer 1 · 2018-12-06T04:22:45.617

0

You can do census.groupby(level=0).nlargest(3)

Thanks @coldspeed

edited Dec 06 '18 at 04:22

answered Dec 06 '18 at 03:40

Axel Puig

`AttributeError: 'SeriesGroupBy' object has no attribute 'high'` what is this, is this actually working for you? – cs95 Dec 06 '18 at 03:57
No, that's still wrong? That doesn't give you the largest observations either. – cs95 Dec 06 '18 at 04:21
Indeed, I was using a sorted dataframe. – Axel Puig Dec 06 '18 at 04:24

1 Answers1