DOC: Remove mention that on with MultiIndex is supported with Rolling Groupby

[GeorgievF]

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

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

---

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

dates = pd.date_range(start='2016-01-01 09:30:00', periods=20, freq='s')
df = pd.DataFrame({'A': [1] * 20 + [2] * 12 + [3] * 8,
                   'B': np.concatenate((dates, dates)),
                   'C': np.arange(40),})
df.set_index(['A', 'B'], inplace=True)
df.groupby('A').rolling('4s', on='B').C.mean()

Problem description

Multi-index dataframe gives the following error message:

ValueError: invalid on specified as B, must be a column (of DataFrame), an Index or None

when one layer of the multi index is used for groupby, while the other is used as a rolling datetime level.

Similar to #13966

Expected Output

A  B                  
1  2016-01-01 09:30:00     0.0
   2016-01-01 09:30:01     0.5
   2016-01-01 09:30:02     1.0
   2016-01-01 09:30:03     1.5
   2016-01-01 09:30:04     2.5
   2016-01-01 09:30:05     3.5
   2016-01-01 09:30:06     4.5
   2016-01-01 09:30:07     5.5
   2016-01-01 09:30:08     6.5
   2016-01-01 09:30:09     7.5
   2016-01-01 09:30:10     8.5
   2016-01-01 09:30:11     9.5
   2016-01-01 09:30:12    10.5
   2016-01-01 09:30:13    11.5
   2016-01-01 09:30:14    12.5
   2016-01-01 09:30:15    13.5
   2016-01-01 09:30:16    14.5
   2016-01-01 09:30:17    15.5
   2016-01-01 09:30:18    16.5
   2016-01-01 09:30:19    17.5
2  2016-01-01 09:30:00    20.0
   2016-01-01 09:30:01    20.5
   2016-01-01 09:30:02    21.0
   2016-01-01 09:30:03    21.5
   2016-01-01 09:30:04    22.5
   2016-01-01 09:30:05    23.5
   2016-01-01 09:30:06    24.5
   2016-01-01 09:30:07    25.5
   2016-01-01 09:30:08    26.5
   2016-01-01 09:30:09    27.5
   2016-01-01 09:30:10    28.5
   2016-01-01 09:30:11    29.5
3  2016-01-01 09:30:12    32.0
   2016-01-01 09:30:13    32.5
   2016-01-01 09:30:14    33.0
   2016-01-01 09:30:15    33.5
   2016-01-01 09:30:16    34.5
   2016-01-01 09:30:17    35.5
   2016-01-01 09:30:18    36.5
   2016-01-01 09:30:19    37.5
Name: C, dtype: float64

Output of pd.show_versions()

INSTALLED VERSIONS

commit : 9d598a5
python : 3.8.5.final.0
python-bits : 64
OS : Windows
OS-release : 10
Version : 10.0.19041
machine : AMD64
processor : Intel64 Family 6 Model 42 Stepping 7, GenuineIntel
byteorder : little
LC_ALL : None
LANG : None
LOCALE : English_United Kingdom.1252

pandas : 1.2.1
numpy : 1.19.2
pytz : 2021.1
dateutil : 2.8.1
pip : 20.3.3
setuptools : 52.0.0.post20210125
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 2.11.3
IPython : 7.20.0
pandas_datareader: None
bs4 : None
bottleneck : None
fsspec : None
fastparquet : None
gcsfs : None
matplotlib : 3.3.2
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pyxlsb : None
s3fs : None
scipy : 1.6.0
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : 1.2.0
xlwt : None
numba : None

[mroeschke]

Thanks for the report. Currently rolling on a multiindex level is unsupported: #34642

Going to close this issue as a duplicate of the linked issue.

[GeorgievF]

It is described in the documents as a possibility. If the documentation is changed that wouldn't be a issue :)
By the way, I was asked in the linked issue to open a new issue for this particular situation.

[mroeschke]

Good point; it is still mentioned in the documentation as a possibility. I will reopen this issue to note to correct the documentation.