BUG: SeriesGroupBy.value_counts - index name missing when applied on categorical column

[donok1]

• 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 master branch of pandas.

Reproducible Example

import pandas as pd
df = pd.DataFrame({'gender': ['female', 'male', 'female', 'male', 'female', 'male'],
                  'education': ['low', 'medium', 'high', 'low', 'high', 'low'],
                  'country': ['US', 'FR', 'US', 'FR', 'FR', 'US']})
df.dtypes  
gender       object
education    object
country      object
dtype: object

df.groupby('country')['gender'].value_counts()

country  gender
FR       male      2
         female    1
US       female    2
         male      1
Name: gender, dtype: int64

df.groupby('country')['gender'].value_counts().index.names

FrozenList(['country', 'gender'])

Now, if using a categorical type on the genre column, the column name is not saved in the index:

df['gender'] = df['gender'].astype('category')
df.dtypes

gender       category
education      object
country        object
dtype: object

df.groupby('country')['gender'].value_counts()

country
FR       male      2
         female    1
US       female    2
         male      1
Name: gender, dtype: int64

df.groupby('country')['gender'].value_counts().index.names

FrozenList(['country', None])

Issue Description

The column name (here gender) is dropped when applying value_counts() on a column of type category.

It might be link with other problems in handling categorical types: #44001

This was working as of pandas v 1.2.3, and I could reproduce the bug starting with v 1.3.0, still present in v 1.3.4

Expected Behavior

Column name should be added in the index for uniformity with other types, as in the first example above with object type column.

Installed Versions

pd.show_versions() INSTALLED VERSIONS ------------------ commit : 945c9ed python : 3.9.6.final.0 python-bits : 64 OS : Darwin OS-release : 21.1.0 Version : Darwin Kernel Version 21.1.0: Wed Oct 13 17:33:23 PDT 2021; root:xnu-8019.41.5~1/RELEASE_X86_64 machine : x86_64 processor : i386 byteorder : little LC_ALL : None LANG : None LOCALE : None.UTF-8

pandas : 1.3.4
numpy : 1.21.4
pytz : 2021.3
dateutil : 2.8.2
pip : 21.3.1
setuptools : 57.0.0
Cython : None
pytest : 6.2.4
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : 4.6.3
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 3.0.1
IPython : 7.26.0
pandas_datareader: None
bs4 : 4.9.3
bottleneck : None
fsspec : None
fastparquet : None
gcsfs : None
matplotlib : 3.4.3
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pyxlsb : None
s3fs : None
scipy : 1.7.1
sqlalchemy : None
tables : None
tabulate : 0.8.9
xarray : 0.19.0
xlrd : None
xlwt : None
numba : None

[NumberPiOso]

take

[NumberPiOso]

I found this minimal example

In [1]: import pandas as pd
   ...: 
   ...: df = pd.DataFrame(
   ...:     {
   ...:         "gender": ["female"],
   ...:         "country": ["US"],
   ...:     }
   ...: )
   ...: 
   ...: print(df.groupby("country")["gender"].value_counts())
   ...: df["gender"] = df["gender"].astype("category")
   ...: print("*" * 60)
   ...: print(df.groupby("country")["gender"].value_counts())
country  gender
US       female    1
Name: gender, dtype: int64
************************************************************
country        
US       female    1
Name: gender, dtype: int64

[NumberPiOso]

In particular the problem is at the function value_counts in the pandas/core/groupby/generic.py.

In PR #38796 the following code has been introduced:[

pandas/pandas/core/groupby/generic.py

Lines 614 to 621 in 011f116

	if bins is not None:
	if not np.iterable(bins):
	# scalar bins cannot be done at top level
	# in a backward compatible way
	return apply_series_value_counts()
	elif is_categorical_dtype(val.dtype):
	# GH38672
	return apply_series_value_counts()

If I simply delete this code the function will solve our problem. Producing the following output:

In [1]: import pandas as pd
   ...: 
   ...: df = pd.DataFrame(
   ...:     {
   ...:         "gender": ["female"],
   ...:         "country": ["US"],
   ...:     }
   ...: )
   ...: 
   ...: print(df.groupby("country")["gender"].value_counts())
   ...: df["gender"] = df["gender"].astype("category")
   ...: print("*" * 60)
   ...: print(df.groupby("country")["gender"].value_counts())
country  gender
US       female    1
Name: gender, dtype: int64
************************************************************
country  gender    <-------------------------------------------
US       female    1
Name: gender, dtype: int64

However we would fall back into the problem solved at PR #38796.

I will investigate further solutions.