[BUG] Index of empty objects not set in Groupby.apply

[galipremsagar]

Describe the bug
When Groupby.apply is invoked on objects that are empty, their .index objects aren't being set to the grouping keys.

Steps/Code to reproduce bug

In [1]: import pandas as pd

In [2]: import cudf

In [3]: df = cudf.DataFrame({"a": [], "b": [], "c": []}, dtype='datetime64[ns]')
df
In [4]: gb = df.groupby("a")['b']

In [7]: pg = df.to_pandas().groupby("a")['b']

In [9]: pg.apply("idxmin").index
Out[9]: DatetimeIndex([], dtype='datetime64[ns]', name='a', freq=None)

In [10]: pg.apply("idxmin")
Out[10]: Series([], Name: b, dtype: datetime64[ns])

In [11]: gb.apply("idxmin")
Out[11]: Series([], Name: b, dtype: datetime64[ns])

In [12]: gb.apply("idxmin").index
Out[12]: RangeIndex(start=0, stop=0, step=1)

Expected behavior

In [12]: gb.apply("idxmin").index
Out[12]: DatetimeIndex([], dtype='datetime64[ns]', name='a', freq=None)

Environment overview (please complete the following information)

• Environment location: [Bare-metal]
• Method of cuDF install: [from source]