[Python] Use saved pandas metadata to determine default timestamp_as_object in to_pandas()
- Dominant language
- C++
- Stars
- 17.1k
- Forks
- 4.3k
- Avg merge
- 3d 18h
- Merged PRs (30d)
- 91
Description
If there is pandas metadata present and it indicates object dtype, we should use that to avoid conversion to `datetime64[ns]`, but keep `datetime` objects in `to_pandas`.
**Reporter**: [Alenka Frim](https://issues.apache.org/jira/browse/ARROW-18175) / @AlenkaF
#### Related issues:
- [[Python] Year 2263 or later datetimes get mangled when written using pandas](https://github.com/apache/arrow/issues/24959) (is related to)
**Note**: *This issue was originally created as [ARROW-18175](https://issues.apache.org/jira/browse/ARROW-18175). Please see the [migration documentation](https://github.com/apache/arrow/issues/14542) for further details.*
Contributor guide
Research direction
Begin at the Python to_pandas() entry point and trace how pandas metadata is read when timestamp_as_object is chosen. Check the related datetime issue for context, then identify the relevant conversion tests. Done means metadata indicating object dtype preserves datetime objects instead of converting them to datetime64[ns].
Written by the indexing model from the issue text.
Assessment
- Tech stack
- pandas, python
- Domain
- data
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100