BUG: first() loses the timezone in groupby

[chrisaycock]

need a systematic test of all groupby methods here.

This works as expected:

pd.DataFrame({'time': [pd.Timestamp('2012-01-01 13:00:00+00:00')],
              'A': [3]}).groupby('A', as_index=False).head(1)

However, this loses the timezone:

pd.DataFrame({'time': [pd.Timestamp('2012-01-01 13:00:00+00:00')],
              'A': [3]}).groupby('A', as_index=False).first()

[jreback]

so this falls down: https://github.com/pandas-dev/pandas/blob/master/pandas/core/groupby.py#L1929

essentially .values is a DatetimeIndex at this point, not a numpy array. (because its passed block.values higher up). So this needs to be handled.

[mroeschke]

I believe the solution to this issue xrefs to #19026 (here), where np.asarray will convert the datetimetz dtype to just a datetime dtype.