BUG: ValueError on groupby with categoricals

[LukasGelbmann]

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

In specific situations involving categorical columns, a groupby() on two or more columns runs into an error:

Code Sample

import pandas as pd
col = pd.Categorical([0, 1])
df = pd.DataFrame({'A': col, 'B': col, 'C': col})
grouped = df.groupby(['A', 'B']).first()
# ValueError: Shape of passed values is (4, 1), indices imply (2, 1)

Expected Output

Expected output would be something like:

>>> grouped
       C
A B     
0 0    0
  1  NaN
1 0  NaN
  1    1

Instead an exception is thrown, with the error message shown above.

[rhshadrach]

Thanks for reporting this - I get the same exception on master. Computing this as a Series via df.groupby(['A', 'B']).C.first() works and gives the correct result.

[LucasG0]

take