BUG: Series.count() raises exception after upgrading from v0.24.1 to v0.25.3 if use_inf_as_na is enabled for a DateTime series.

[thy09]

Code Sample, a copy-pastable example if possible

# Your code here
import pandas as pd
from datetime import datetime

if __name__ == '__main__':
    s = pd.Series([datetime.now()])
    with pd.option_context('use_inf_as_na', True):
        s.count()

Problem description

Previously with pandas v0.24.1, the above code works well,
however, after I upgrade the version to v0.25.3,
the above code raises an exception:
AttributeError: 'DatetimeArray' object has no attribute '_constructor'

It seems that something goes wrong when passing a DatetimeArray to _isna_old.
I think it is a bug introduced in some recent update.

Expected Output

Output of pd.show_versions()

[paste the output of pd.show_versions() here below this line]

INSTALLED VERSIONS

commit : None
python : 3.6.8.final.0
python-bits : 64
OS : Windows
OS-release : 10
machine : AMD64
processor : Intel64 Family 6 Model 85 Stepping 4, GenuineIntel
byteorder : little
LC_ALL : None
LANG : None
LOCALE : None.None

pandas : 0.25.3
numpy : 1.17.3
pytz : 2019.3
dateutil : 2.8.0
pip : 18.1
setuptools : 40.6.2
Cython : None
pytest : 4.3.0
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 2.10.3
IPython : 7.9.0
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
gcsfs : None
lxml.etree : None
matplotlib : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : 0.12.1
pytables : None
s3fs : None
scipy : 1.2.1
sqlalchemy : None
tables : None
xarray : None
xlrd : None
xlwt : None
xlsxwriter : None

[jorisvandenbossche]

@thy09 late reply, but thanks for the report! I can confirm this is a regression compared to 0.24

I think the fix should be to add ExtensionArray to this isinstance check here:

pandas/pandas/core/dtypes/missing.py

Line 178 in 35029d2

elif isinstance(obj, (ABCSeries, np.ndarray, ABCIndexClass)):

You're certainly welcome to look into it and do a PR!

[thy09]

In _isna_new, I see 3 more types ABCExtensionArray, ABCDatetimeArray, ABCTimedeltaArray,,
should we follow the same logic and add these three types to isna_old?

[jorisvandenbossche]

Normally, if something is a DatetimeArray or TimedeltaArray, it is also an ExtensionArray (it are subclasses), so I would think that only adding ABCExtensionArray would be enough. But you might need to test that.