BUG: groupby.cumcount returns NaN and dtype=float if groupby column contains np.nan

[mluck]

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
import numpy as np

pd.DataFrame({'data':[np.nan,1,1,3,np.nan]}).groupby('data').cumcount(ascending=True)

Issue Description

I'm not sure if this a known or expected behavior, but pandas.DataFrame.groupby.cumcount() returns a float and NaN values when the groupby column contains np.nan. This is changed behavior from past versions (compared with v.1.1.0) where the count of NaN values was returned as an int. In both the older and current version, ints are returned if no NaNs are present, but the dtype returned in v.2.0.0 depends on whether or not there is a NaN.

Expected Behavior

I think the expected behavior would be as in past versions, where all values returned are ints, for example:

>>> pd.__version__
'1.1.0'
>>> pd.DataFrame({'data':[np.nan,1,1,3,np.nan]}).groupby('data').cumcount(ascending=True)
0    0
1    0
2    1
3    0
4    1
dtype: int64

Installed Versions

/usr/local/lib/python3.8/dist-packages/_distutils_hack/init.py:33: UserWarning: Setuptools is replacing distutils.
warnings.warn("Setuptools is replacing distutils.")

INSTALLED VERSIONS

commit : 478d340
python : 3.8.10.final.0
python-bits : 64
OS : Linux
OS-release : 5.15.0-1034-aws
Version : #38~20.04.1-Ubuntu SMP Wed Mar 29 19:48:16 UTC 2023
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : C.UTF-8
LANG : C.UTF-8
LOCALE : en_US.UTF-8

pandas : 2.0.0
numpy : 1.23.5
pytz : 2023.3
dateutil : 2.8.2
setuptools : 67.6.1
pip : 20.0.2
Cython : 0.29.34
pytest : 7.3.0
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : 2.9.6
jinja2 : 3.1.2
IPython : 8.12.0
pandas_datareader: None
bs4 : 4.12.2
bottleneck : None
brotli : 1.0.9
fastparquet : None
fsspec : None
gcsfs : None
matplotlib : 3.7.1
numba : 0.56.4
numexpr : 2.8.4
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : 11.0.0
pyreadstat : None
pyxlsb : None
s3fs : None
scipy : 1.10.1
snappy : None
sqlalchemy : None
tables : 3.8.0
tabulate : None
xarray : 2023.1.0
xlrd : None
zstandard : None
tzdata : 2023.3
qtpy : None
pyqt5 : None

[rhshadrach]

Thanks for the report. This change was highlighted here: https://pandas.pydata.org/docs/whatsnew/v1.5.0.html#using-dropna-true-with-groupby-transforms

groupby has a dropna argument that defaults to True. As your example highlights, this argument was not being respected, and now it is. You can get the desired output by setting dropna=False:

pd.DataFrame({'data':[np.nan,1,1,3,np.nan]}).groupby('data', dropna=False).cumcount(ascending=True)
# 0    0
# 1    0
# 2    1
# 3    0
# 4    1
# dtype: int64

Does this resolve your issue?

[mluck]

Yes, that's exactly what I needed.

Thanks so much!