BUG: Fix read_hdf failing on generic datetime64 dtype #64006

[sanrishi]

• closes BUG: pd.read_hdf unable to retrieve a pd.Series with dtype as "datetime64" since v3.0.0 #64006 (Replace xxxx with the GitHub issue number)
• Tests added and passed 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.
• I have reviewed and followed all the contribution guidelines
• If I used AI to develop this pull request, I prompted it to follow AGENTS.md.

Summary of Changes

This PR fixes a regression in read_hdf where reading HDF5 files created by older pandas versions (containing datetime64 metadata without an explicit unit) would raise a TypeError: 'generic' is not a valid TimeUnit.

Implementation Details

• Added a check in pandas/io/pytables.py (in both restore_kwargs and read_array) to handle the bare string "datetime64".

• If the dtype string is exactly "datetime64" (generic), it now defaults to "datetime64[ns]".

• This mirrors the existing legacy handling logic for timedelta64 found immediately adjacent to the fix.

Verification

Since generating "broken" legacy files with modern pandas/numpy is difficult (as they now enforce units), I added a regression test in pandas/tests/io/pytables/test_store.py that uses h5py to manually strip unit metadata from a test file, simulating the legacy format.

Tests Added:

1. test_read_hdf_datetime64_without_unit_gh64006: Verifies that generic datetime64 metadata is correctly read as [ns].

2. test_read_hdf_preserves_explicit_units: Verifies that files with explicit units (e.g., [s], [ms]) are not overwritten by this fix.

(pandas-dev) C:\Users\My\Documents\GitHub\pandas>pytest pandas/tests/io/pytables/test_store.py -k "gh64006 or datetime"
C:\Users\My\.conda\envs\pandas-dev\Lib\site-packages\pytest_cython\__init__.py:2: UserWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html. The pkg_resources package is slated for removal as early as 2025-11-30. Refrain from using this package or pin to Setuptools<81.
  from pkg_resources import get_distribution
=========================================================================================================== test session starts ============================================================================================================
platform win32 -- Python 3.11.14, pytest-9.0.2, pluggy-1.6.0
PySide6 6.9.3 -- Qt runtime 6.9.3 -- Qt compiled 6.9.3
rootdir: C:\Users\My\Documents\GitHub\pandas
configfile: pyproject.toml
plugins: anyio-4.12.0, hypothesis-6.148.7, cov-7.0.0, cython-0.3.1, localserver-0.0.0, qt-4.5.0, xdist-3.8.0
collected 78 items / 72 deselected / 6 selected

pandas\tests\io\pytables\test_store.py s....s

---------------------------------------------------------------------------------- generated xml file: C:\Users\My\Documents\GitHub\pandas\test-data.xml -----------------------------------------------------------------------------------
=========================================================================================================== slowest 30 durations ===========================================================================================================
0.04s call     pandas/tests/io/pytables/test_store.py::test_read_hdf_datetime_units_preserved[s]
0.03s setup    pandas/tests/io/pytables/test_store.py::test_read_hdf_datetime64_without_unit_gh64006
0.02s call     pandas/tests/io/pytables/test_store.py::test_read_hdf_datetime_units_preserved[ns]
0.02s call     pandas/tests/io/pytables/test_store.py::test_read_hdf_datetime_units_preserved[us]
0.02s call     pandas/tests/io/pytables/test_store.py::test_read_hdf_datetime_units_preserved[ms]

(13 durations < 0.005s hidden.  Use -vv to show these durations.)
=============================================================================================== 4 passed, 2 skipped, 72 deselected in 1.12s ================================================================================================

[sanrishi]

Pre-commit.ci autofix