groupby - transform not working with multiple columns

[behzadnouri]

I came across this issue while looking at the this question on stackoverflow;

say I have a data-frame like:

>>> df
  tag  val1  val2
0   B     0     0
1   A     1     2
2   A     2     4
3   B     3     6
4   B     4     8

and I want to reverse columns val1 and val2 within each tag; this works like a charm:

>>> grb = df.groupby('tag')
>>> df['val1'] = grb['val1'].transform(lambda ts: ts.loc[::-1])
>>> df['val2'] = grb['val2'].transform(lambda ts: ts.loc[::-1])
>>> df
  tag  val1  val2
0   B     4     8
1   A     2     4
2   A     1     2
3   B     3     6
4   B     0     0

but, if I want to do this at the same time for columns val1 and val2, it breaks:

>>> grb = df.groupby('tag')
>>> grb['val1', 'val2'].transform(lambda obj: obj.loc[::-1])
   val1  val2
0     0     0
1     1     2
2     2     4
3     3     6
4     4     8

[5 rows x 2 columns]

obviously I will not get the correct answer if I overwrite df with the new data ( or append the new columns ); however, this

>>> grb = df.groupby('tag')
>>> grb['val1', 'val2'].transform(lambda obj: obj.values[::-1])
   val1  val2
0     4     8
1     2     4
2     1     2
3     3     6
4     0     0

[5 rows x 2 columns]

works as expected; which makes me guess this has something to do with the index being aligned with the original data-frame, whereas in here it should be ignored;

[jreback]

this might be fixed by this: #5264

if not, pls submit a PR (its just a small fix I think)

[jreback]

this is not a bug and works as expected