BUG: drop_duplicates on categorical not preserving NA

[phofl]

• 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

ser = Series(
    Categorical(
        [True, False, True, False, NA], categories=[True, False], ordered=True
    )
)
ser.drop_duplicates()

Issue Description

NA gets converted to NaN

0     True
1    False
4      NaN
dtype: category
Categories (2, object): [True < False]

Expected Behavior

Returning NA

0     True
1    False
4       NA
dtype: category
Categories (2, object): [True < False]

Installed Versions

INSTALLED VERSIONS ------------------ commit : 01b86ed python : 3.8.12.final.0 python-bits : 64 OS : Linux OS-release : 5.11.0-38-generic Version : #42~20.04.1-Ubuntu SMP Tue Sep 28 20:41:07 UTC 2021 machine : x86_64 processor : x86_64 byteorder : little LC_ALL : None LANG : en_US.UTF-8 LOCALE : en_US.UTF-8

pandas : 1.4.0.dev0+1085.g01b86edbbb
numpy : 1.21.3
pytz : 2021.1
dateutil : 2.8.2
pip : 21.2.4
setuptools : 58.0.4
Cython : 0.29.24
pytest : 6.2.5
hypothesis : 6.23.1
sphinx : 4.2.0
blosc : None
feather : None
xlsxwriter : 3.0.1
lxml.etree : 4.6.3
html5lib : 1.1
pymysql : None
psycopg2 : None
jinja2 : 3.0.1
IPython : 7.28.0
pandas_datareader: None
bs4 : 4.10.0
bottleneck : 1.3.2
fsspec : 2021.11.0
fastparquet : 0.7.1
gcsfs : 2021.05.0
matplotlib : 3.4.3
numexpr : 2.7.3
odfpy : None
openpyxl : 3.0.9
pandas_gbq : None
pyarrow : 5.0.0
pyxlsb : None
s3fs : 2021.11.0
scipy : 1.7.2
sqlalchemy : 1.4.25
tables : 3.6.1
tabulate : 0.8.9
xarray : 0.18.0
xlrd : 2.0.1
xlwt : 1.3.0
numba : 0.53.1
None

Process finished with exit code 0

[jorisvandenbossche]

Note that this is not actually related to drop_duplicated. Also the initial categorical series you created will show "NaN" in the repr.

Missing values in Categorical are stored custom as -1 in the integer codes (and not in the categories, so not depending on the actual data type of the categories).
And we don't yet have support in general for "nullable" Categorical (i.e. using NA and NA-semantics for their missing values, and allowing to store a nullable array for the categories).

[phofl]

Should we relabel as enhancement for general nullable categorical? Had a quick look unter NA - MaskerArrays and have not seen a duplicate

[simonjayhawkins]

Should we relabel as enhancement for general nullable categorical? Had a quick look unter NA - MaskerArrays and have not seen a duplicate

I think covered by #43836 and #29962. closing to help discussion in less places.