Pandas Groupby apply function to count values greater than zero

Question

I am using groupby and agg in the following manner:

df.groupby('group')['a'].agg({'mean' : np.mean, 'std' : np.std})

and I would like to also count the values above zero in the same column ['a']

the following line does the count as I want,

sum(x > 0 for x in df['a'])

but I can't get it work when applying to groupby.

Following an example for applying a pandas calculation to a groupby I tried:

df.groupby('group')['a'].apply(sum(x > 0 for x in df['a']))

but I get an error message: AttributeError: 'numpy.int32' object has no attribute 'module'

Can anybody please suggest how this might be done?

`.agg({'pos':lambda ts: (ts > 0).sum()})` – behzad.nouri Mar 31 '14 at 00:00 — behzad.nouri, Mar 31 '14 at 00:00
@behzad.nouri you should make that an answer. – Andy Hayden Mar 31 '14 at 06:15 — Andy Hayden, Mar 31 '14 at 06:15

score 14 · Accepted Answer · edited Apr 23 '20 at 02:20

Answer from the comments:

 .agg({'pos':lambda ts: (ts > 0).sum()}) # –  behzad.nouri Mar 31 at 0:00

This is my contribution to the backlog of unanswered questions :) Credits to behzad.nouri

Update 2020 In the latest pandas version, you need to do the following:

 .agg(pos=lambda ts: (ts > 0).sum())

otherwise it will result in the following error:

SpecificationError: nested renamer is not supported

1 Answers1