Closed windiana42 closed 5 months ago
Snowflake support doesn't seem to be that hard for pipedag. However, for small fast tests Snowflake seems to be awefully slow:
This PR is also about fixing compatibility issues with pandas < 2 and sqlalchemy < 2 compatibility. Especially pyarrow handling was tricky. I opted for preferring StringDtype("pyarrow")
over ArrowDtype(pa.string())
. Future will tell how good this choice was. There is no problem to revise it. But it must be consistent in a few places.
Checklist
docs/source/changelog.md
entrydocs/source/
docs/source/examples.md