BUG: Runtime warning with groupby/tail when `None` appears in group column

[ian-r-rose]

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
df = pandas.DataFrame(
    [
        ["a", 1],
        ["a", 2],
        [None, 0],
        ["b", 2],
        ["b", 3],

    ],
    columns=["x", "y"],
)
df.groupby("x", dropna=False).tail(1)   # Succeeds
df.groupby("x", dropna=True).tail(1)   # Produces warning

Issue Description

👋 In pandas main, a groupby/tail on a DataFrame which contains nulls in the grouped column produces a RuntimeWarning suggesting a mishandled case in indexing logic:

/lib/python3.8/site-packages/pandas/core/groupby/indexing.py:217: RuntimeWarning: invalid value encountered in remainder
  mask &= offset_array % step == 0

If I set dropna=False, the snippet succeeds without producing a warning.

Based on the warning location and git blame, it may be related to #42947

Expected Behavior

The above groupby should succeed without producing a warning.

Installed Versions

Pandas main, as well as 1.4.

[rhshadrach]

Thanks for the report! I'm not seeing this on main or 1.4.x; can you post the output of pd.show_versions().

versions

pandas           : 1.4.2+11.gd1717e98a2
numpy            : 1.22.3
pytz             : 2022.1
dateutil         : 2.8.2
pip              : 22.0.4
setuptools       : 62.1.0
Cython           : 0.29.28
pytest           : 7.1.1
hypothesis       : 6.43.3
sphinx           : 4.5.0
blosc            : None
feather          : None
xlsxwriter       : 3.0.3
lxml.etree       : 4.8.0
html5lib         : 1.1
pymysql          : None
psycopg2         : None
jinja2           : 3.0.3
IPython          : 8.2.0
pandas_datareader: None
bs4              : 4.11.1
bottleneck       : 1.3.4
brotli           : 
fastparquet      : 0.8.0
fsspec           : 2021.11.0
gcsfs            : 2021.11.0
markupsafe       : 2.1.1
matplotlib       : 3.5.1
numba            : 0.53.1
numexpr          : 2.8.0
odfpy            : None
openpyxl         : 3.0.9
pandas_gbq       : None
pyarrow          : 3.0.0
pyreadstat       : 1.1.4
pyxlsb           : None
s3fs             : 2021.11.0
scipy            : 1.8.0
snappy           : 
sqlalchemy       : 1.4.35
tables           : 3.7.0
tabulate         : 0.8.9
xarray           : 0.18.2
xlrd             : 2.0.1
xlwt             : 1.3.0
zstandard        : None

[ian-r-rose]

Ah, my mistake, I can't reproduce on 1.4.x, so I must have messed up my test environment yesterday (sorry to omit show_versions). I can still reproduce this on main:

pd.show_versions():

INSTALLED VERSIONS ------------------ commit : a8968bf python : 3.8.8.final.0 python-bits : 64 OS : Linux OS-release : 5.13.0-39-generic Version : #44~20.04.1-Ubuntu SMP Thu Mar 24 16:43:35 UTC 2022 machine : x86_64 processor : x86_64 byteorder : little LC_ALL : None LANG : en_US.UTF-8 LOCALE : en_US.UTF-8

pandas : 1.5.0.dev0+694.ga8968bfa69
numpy : 1.21.5
pytz : 2021.3
dateutil : 2.8.2
pip : 21.3.1
setuptools : 59.8.0
Cython : 0.29.22
pytest : 6.2.2
hypothesis : None
sphinx : 4.4.0
blosc : 1.10.2
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : 1.1
pymysql : None
psycopg2 : None
jinja2 : 2.11.3
IPython : 7.30.1
pandas_datareader: None
bs4 : 4.10.0
bottleneck : None
brotli :
fastparquet : 0.7.2
fsspec : 2021.06.1
gcsfs : 0.7.2
markupsafe : 1.1.1
matplotlib : 3.3.4
numba : 0.53.1
numexpr : 2.8.0
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : 5.0.0
pyreadstat : None
pyxlsb : None
s3fs : 2021.06.1
scipy : 1.7.3
snappy :
sqlalchemy : 1.4.20
tables : 3.6.1
tabulate : 0.8.9
xarray : 0.21.1
xlrd : None
xlwt : None
zstandard : None

[ian-r-rose]

Update: I ran a git bisect, and for me the warning starts to show up at 4d7a03a

[rhshadrach]

Thanks, I'm able to reproduce with numpy 1.21.6; no warning appears with numpy 1.22.0.This was a bug that fixed in 1.22.0:

numpy/numpy#18170

We could add a workaround to avoid the warning for numpy < 1.22 by computing a Boolean mask where the array is null.

[ian-r-rose]

Thanks @rhshadrach! Upgrading numpy is a fine workaround from my perspective.