Pandas BUG: groupby.std with no numeric columns and numeric_only=True raises

df = pd.DataFrame({'a': list('xyz'), 'b': list('def')})
gb = df.groupby('a')
print(gb.std(numeric_only=True))

raises TypeError: All columns were dropped in grouped_reduce. Instead, it should return an empty frame with index consisting of the groups.

Comment From: rhshadrach

I haven't verified this, but it looks like this was fixed in

commit bcb6c1adf5f18465389dad781436626dddd6c64b
Author: jbrockmendel
Date:   Tue Jan 31 12:03:29 2023 -0800

    CLN: assorted (#51078)

cc @jbrockmendel. Could use tests.

Comment From: NumanIjaz

take