Pandas groupby().sem reorders by columns after adding a non-numeric column

Asked Dec 27 '16 at 16:09

Active Dec 27 '16 at 23:17

Viewed 170 times

first time asking a question: I'm using

print(df.groupby(['SIDE']).sem())

get correct results when "Sector" column, the last in the data set has a heading but no data but after populating non-numeric data in the column it returns data frame with columns in different order. I'm also using

df.groupby('SIDE').mean()
df.groupby('SIDE').std()

with no issues on the same data, feeling stuck, seems like a bug?

enter image description here

After adding non-numeric data to column

edited Dec 27 '16 at 23:17

piRSquared

265,629
48
427
571

asked Dec 27 '16 at 16:09

Andrew Bunker

2

I'm confused! Please read [MCVE](http://stackoverflow.com/help/mcve) – piRSquared Dec 27 '16 at 16:15
Also [how to make a reproducible example with pandas](http://stackoverflow.com/questions/20109391/how-to-make-good-reproducible-pandas-examples). – lmo Dec 27 '16 at 16:19
@AndrewBunker Thou shalt not post images. Thou shalt post it as text! – Mohammad Yusuf Dec 27 '16 at 16:30
Yes, that looks like a bug. A workaround would be `df.groupby(['SIDE']).sem()[df.columns]` – ayhan Dec 27 '16 at 17:31
1

Thanks guys for the help and intro to the posting protocol. Ayhan, your post helped, I finally used the reindex_like function to work around. – Andrew Bunker Dec 27 '16 at 23:27

Pandas groupby().sem reorders by columns after adding a non-numeric column

0 Answers0