- add polars backend
- add Date and Duration type
- string / datetime operations now have their separate namespace (.str / .dt)
- add partition_by=, arrange= and filter= arguments for window / aggregation functions (filter does not work on SQL yet)
- migrate project to pixi
- support ArrowDtype based columns in pandas dataframes
- better support for apache arrow backed pandas dataframes
- fix handling of boolean literals
- fix literal handling within SQL expressions
- support for operators/functions that take constant arguments
- support pandas dataframes backed by pyarrow extension dtypes
- relax python version requirement (>=3.9, <4).
- development of pydiverse.transform is currently slow since pydiverse.pipedag adoption is currently prioritized: this will help to limit scope that then can be provided comprehensively with high quality
- added support for pandas >= 2.0.0
- added support for sqlalchemy >= 2.0.0
- Initial release.