pandas (original) (raw)

closes ENH: Support for non-nanosecond resolution in read_csv and to_datetime #54048
closes BUG: Inconsistent datetime conversion behavior when constructing a DataFrame with Python datetimes. #55014
closes ENH: Support non-nano inference in to_datetime #55096
closes BUG: Assignment of Timestamp Scalar uses micrsosecond precision, Series uses nano #55487
closes ENH/API: resolution inference in vectorized datetime parsing #55564
closes BUG: Inconsistent datetime dtype precision with assign() using list vs direct value #56733
closes BUG: Inconsistent resolution from constructing with and assigning datetime values #57080
closes BUG: Out of bounds nanosecond timestamp when s was specified #57349
closes BUG: Series constructor loses date unit of Timestamps #58300
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.

This is an API change so will have to wait for 3.0, posting it to get feedback on desired API in mostly-IO places where this causes API changes.

This does not yet handle Timedelta scalar, TimedeltaIndex, or pd.to_timedelta. I'd like to get those done for 3.0 but its a much lower priority than the datetime side.

I've been splitting things off to reduce the diff, will try to keep doing that. Open to suggestions for splittable components. Currently #55741 has some of the cython implementation.

19 tests are failing locally. 11 of those are stata tests that would be addressed by #55642. 4 are SAS and have a similar option available. One is in JSON and may be hard to solve, xref #55827. The rest are caused by #55737.