BUG: Categorical series .isin({pd.NaT, ...}) gives all-False

[adamhooper]

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

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

• (optional) I have confirmed this bug exists on the master branch 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

import pandas as pd
pd.Series([pd.NaT, ''], dtype='category').isin({pd.NaT, ''})  # [True, True]
pd.Series([''], dtype='category').isin({''})  # [True] because '' is in series
pd.Series([''], dtype='category').isin({None, ''})  # [True] because '' is in series
pd.Series([''], dtype='category').isin({pd.NaT, ''})  # [False] even though '' is in series

Problem description

When searching for empty values in a categorical full of empty values, searching for pd.NaT will prevent other terms from being found.

(Perhaps that's because Pandas casts all input values to Timedelta?)

This behavior is unexpected. It makes me feel like I must be crazy. If I'm not meant to search for pd.NaT, I'd feel very happy with an error message when I try this.

Expected Output

pd.Series([''], dtype='category').isin({pd.NaT, ''})  # [True] because '' is in series

Output of pd.show_versions()

INSTALLED VERSIONS

commit : 2a7d332
python : 3.8.5.final.0
python-bits : 64
OS : Linux
OS-release : 5.8.4-200.fc32.x86_64
Version : #1 SMP Wed Aug 26 22:28:08 UTC 2020
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.1.2
numpy : 1.19.2
pytz : 2019.3
dateutil : 2.8.0
pip : 19.3.1
setuptools : 45.2.0
Cython : None
pytest : 5.4.3
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : 4.4.1
html5lib : 1.0.1
pymysql : None
psycopg2 : None
jinja2 : 2.11.2
IPython : None
pandas_datareader: None
bs4 : 4.9.1
bottleneck : None
fsspec : None
fastparquet : None
gcsfs : None
matplotlib : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : 0.16.0
pytables : None
pyxlsb : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
numba : None

[dsaxton]

@adamhooper Thanks, seems there is a bug here. The value doesn't have to be pd.NaT, it can be other datetime-like things, so maybe your suspicion is right that the empty string is being converted somewhere (other values don't seem to have the same effect):

[ins] In [1]: import numpy as np
         ...: import pandas as pd
         ...:
         ...:
         ...: value = ""
         ...:
         ...: s = pd.Series([value], dtype="category")
         ...: print(s.isin([pd.Timedelta(0), value]))
         ...: print(pd.__version__)
         ...:
0    False
dtype: bool
1.2.0.dev0+469.gf04624104

[samilAyoub]

take

[topper-123]

This works as expected in main (v2.0rc1):

>>> import pandas as pd
>>>
>>> pd.Series([pd.NaT, ''], dtype='category').isin({pd.NaT, ''})
0    True
1    True
dtype: bool
>>> pd.Series([''], dtype='category').isin({''})
0    True
dtype: bool
>>> pd.Series([''], dtype='category').isin({None, ''})
0    True
dtype: bool
>>> pd.Series([''], dtype='category').isin({pd.NaT, ''})  # this gave [False] before
0    True
dtype: bool

I'll add a test for this so it doesn't pop up again.