check ExtensionType in is_datetime64_any_dtype for array-likes

[jmoralez]

• closes BUG: is_datetime64_any_dtype returns False for Series with pyarrow dtype #57055
• Tests added and passed](https://pandas.pydata.org/pandas-docs/dev/development/contributing_codebase.html#writing-tests) if fixing a bug or adding a new feature
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

[jmoralez]

Failures in CI look unrelated to this PR (logs)

FAILED pandas/tests/arrays/categorical/test_sorting.py::TestCategoricalSort::test_numpy_argsort - Failed: DID NOT RAISE <class 'ValueError'>
FAILED pandas/tests/extension/test_categorical.py::TestCategorical::test_argsort - AssertionError: Series are different

[mroeschke]

Could you add a test in test_arrow.py?

[jmoralez]

Thanks @mroeschke. The test is failing (log) for arrow date types

FAILED pandas/tests/extension/test_arrow.py::test_is_datetime64_any_dtype[date32[day]] - AssertionError
FAILED pandas/tests/extension/test_arrow.py::test_is_datetime64_any_dtype[date64[ms]] - AssertionError

since for those types pa.types.is_timestamp returns False. However the current implementation of is_datetime64_any_dtype returns True when provided as a dtype, e.g. pd.api.types.is_datetime64_any_dtype(serie_with_pyarrow_date32.dtype). What should I do here? Is it expected for that to return True or should I adjust the above condition to also check if pa.types.is_date returns False?

Edit
I checked that both of those types translate to datetime64 in numpy:

import pyarrow as pa

type(pa.date32().to_pandas_dtype())
# numpy.dtypes.DateTime64DType 
type(pa.date64().to_pandas_dtype())
# numpy.dtypes.DateTime64DType

so I've included those in the test for True

[mroeschke]

I think the date types should return False. In pandas they don't always behave like date times

[jmoralez]

Agree, for example trying to extract the second raises an error. I'll make the changes.

[mroeschke]

Thanks @jmoralez