Pandas groupby ewm

Become part of the top 3% of the developers by applying to Toptal https://topt.al/25cXVn

--

Music by Eric Matyas
https://www.soundimage.org
Track title: Lost Civilization

--

Chapters
00:00 Question
02:50 Accepted answer (Score 8)
03:05 Answer 2 (Score 0)
03:26 Answer 3 (Score 0)
03:47 Thank you

--

Full question
https://stackoverflow.com/questions/5800...

Question links:
[SO: Pandas Groupby and apply method with custom function]: https://stackoverflow.com/questions/4950...
[image]: https://i.stack.imgur.com/9hLd8.png

Accepted answer links:
[transform]: https://pandas.pydata.org/pandas-docs/st...

--

Content licensed under CC BY-SA
https://meta.stackexchange.com/help/lice...

--

Tags
#python #pandas #timeseries #pandasgroupby #rollingcomputation

#avk47

ACCEPTED ANSWER

Let's fix the problem, using transform:

t['ewm'] = ts.groupby(['C1'])['X1'].transform(lambda x: x.ewm(halflife=10).mean()).values()

The accepted answer is very slow for large datasets.

What I have done is:

ts['ewm'] = ts.groupby(['C1']).ewm(halflife=10).mean().values

and it seems to work just fine

Can you try this? Do NOT set ts = df.set_index('T'). Then you can do as below

ts['ewm']=ts.groupby(['C1'], sort=False).apply(lambda x: x['X1'].ewm(halflife=10).mean()).reset_index(drop=True)