Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Not A Bug
-
None
-
None
Description
It looks like there was a slowdown in our dataframe-to-table benchmark after Arrow-13164.
I'm a bit surprised since that PR is about arrow -> R conversion (though there might be some crossover).
Here's a recent conbench link that shows the the regression https://conbench.ursa.dev/benchmarks/8533b3e9ea654bcea8ccd1fe448861ba/
It looks like it happens with types: integer, float, and dict, but not strings