BUG: Performing groupby with `as_index=False` when `df.columns.name` is not None drops the name

[sfc-gh-joshi]

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(
    [[1, 2, 3]],
    columns=pd.Index(["A", "B", "C"], name="alpha")
)
df.groupby(by="A", as_index=False)["B"].sum()

Issue Description

Output of the above example:

>>> df
alpha  A  B  C
0      1  2  3
>>> df.groupby(by="A", as_index=False)["B"].sum()
   A  B
0  1  2

The name field of the columns Index object is dropped after the groupby operation.

>>> df.groupby(by="A", as_index=False)["B"].sum().columns
Index(['A', 'B'], dtype='object')

In pandas 1.5, the name was retained:

>>> pd.__version__
'1.5.3'
>>> df.groupby(by="A", as_index=False)["B"].sum()
alpha  A  B
0      1  2
>>> df.groupby(by="A", as_index=False)["B"].sum().columns
Index(['A', 'B'], dtype='object', name='alpha')

Expected Behavior

It is unclear whether this change was intentional between pandas 1.5/2.0; I could not find any mention of this in the issue tracker or release notes. If this was intentional, it should be mentioned briefly in groupby documentation; if not, then it should be fixed to preserve the name as before.

Installed Versions

INSTALLED VERSIONS

commit : 7fe86b6
python : 3.11.7.final.0
python-bits : 64
OS : Darwin
OS-release : 23.4.0
Version : Darwin Kernel Version 23.4.0: Wed Feb 21 21:45:49 PST 2024; root:xnu-10063.101.15~2/RELEASE_ARM64_T6020
machine : arm64
processor : arm
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 3.0.0.dev0+263.g7fe86b67d3
numpy : 1.26.4
pytz : 2024.1
dateutil : 2.8.2
setuptools : 68.2.2
pip : 23.3.1
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : None
IPython : None
pandas_datareader : None
adbc-driver-postgresql: None
adbc-driver-sqlite : None
bs4 : None
bottleneck : None
dataframe-api-compat : None
fastparquet : None
fsspec : None
gcsfs : None
matplotlib : None
numba : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : 15.0.0
pyreadstat : None
python-calamine : None
pyxlsb : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
zstandard : None
tzdata : 2023.4
qtpy : None
pyqt5 : None

[rhshadrach]

Thanks for the report! This was changed in #50744 and should have been noted in a whatsnew. As noted in that PR, by dropping name, we get consistency between df["B"].groupby(by=...).sum() and df.groupby(by=...)["B"].sum() (with or without as_index=False). Also, I think Series selection in groupby should act similar to Series selection for a DataFrame, and as such the current behavior appears correct to me.

Note that if you use [["B"]] instead, the name is preserved (same as doing df[["B"]]).

I don't believe there is a section detailing column selection in the groupby User Guide. I am in favor of a small section on this, and the name could be mentioned there.

[shriyakalakata]

take

[shriyakalakata]

Hey @rhshadrach! I wanted to clarify what's expected from this task. I see that there is a 'column selection in GroupBy' in the groupby user guide already. I was thinking I could do the following:

• Add a note in this section about the name
• Include an example of grouping by a certain column and using another column for aggregation

[rhshadrach]

Thanks @shriyakalakata - indeed I missed this section!

• Add a note in this section about the name

This sounds good.

• Include an example of grouping by a certain column and using another column for aggregation

I don't think this is necessary - there are already examples of this throughout the groupby docs.