df.groupby on a column with complex numbers is broken

[basnijholt]

Currently, grouping by a key that has complex numbers is broken, see the following example

import pandas as pd
df = pd.DataFrame([dict(a=1, b=1 + 1j), dict(a=1, b=1 + 2j)])
df.groupby("b").groups

output:

{1.0: Int64Index([0, 1], dtype='int64')}

expected:

{0: Int64Index([0], dtype='int64'), 1: Int64Index([1], dtype='int64')}

Related #10921

[jschendel]

Probably the same underlying issue as #17927

[rhshadrach]

I'm now seeing {(1+1j): [0], (1+2j): [1]} from the OP. Could likely use tests.

[xiaohuanlin]

Hi @rhshadrach, can I take this question? I think the expection for this question is to create a unittest to prevent it happen again, is it right?

[aram-cinnamon]

take

[aram-cinnamon]

Ah sorry @xiaohuanlin, I already picked up this issue on Sunday (You can refer to our slack channel). Next time I'll take early so there's no confusion.