Pandas version checks

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

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

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

import pandas as pd

df = pd.DataFrame({'x': [10, 20, 30], 'y': ['a', 'b', 'c']})

# group by atomic - names and keys are atomic
name, _ = next(iter(df.groupby('x')))
name    # `10` as expected
df.groupby('x').groups  # `{10: [0], 20: [1], 30: [2]}` as expected

# group by list of 2 - names and keys are 2-tuples
name, _ = next(iter(df.groupby(['x', 'y'])))
name  # `(10, 'a')` as expected
df.groupby(['x', 'y']).groups  # `{(10, 'a'): [0], (20, 'b'): [1], (30, 'c'): [2]}` as expected

# group by list of 1 - names are 1-tuples but keys are atomic!?
name, _ = next(iter(df.groupby(['x'])))
name  # `(10,)` as expected
df.groupby(['x']).groups  # `{10: [0], 20: [1], 30: [2]}` UNEXPECTED

Issue Description

In pandas 1.5.x grouping by a single element yielded scalar group names, grouping by a list of more than one element yielded tuples names but grouping by a length-1 list yielded scalar group names again which was considered inconsistent (and I agree).

This was mostly resolved in 2.0 (see #42795) and now when iterating through groups the group names are 1-tuples when grouped with a length-1 list. However the groups attribute still returns scalars instead of 1-tuples which I consider to be a bug or at least requiring enhancement.

Expected Behavior

I expect group names / keys to be 1-tuples when grouping with a length-1 list: group by scalar get scalar names, group by list get tuple names. (I would also propose that grouping by an empty list should return a single group of the whole dataframe, with an empty tuple as the key, but thats out of scope for this issue!)

I expect the same group name / key behaviour whether iterating through groups or accessing groups via the groups attribute.

Installed Versions