Potential regression induced by PR #57479

[rhshadrach]

PR #57479 may have induced a performance regression. If it was a necessary behavior change, this may have been expected and everything is okay.

Please check the links below. If any ASVs are parameterized, the combinations of parameters that a regression has been detected for appear as subbullets.

• frame_methods.Fillna.time_fillna
  • inplace=True; dtype='float32'
  • inplace=True; dtype='float64'

Subsequent benchmarks may have skipped some commits. The link below lists the commits that are between the two benchmark runs where the regression was identified.

Commit Range

cc @phofl

[rhshadrach]

Not certain this is the PR that introduced the regression. Is it possible that making the shallow copy improves performance?

[DeaMariaLeon]

@phofl This is the commit I mentioned to you on Slack (the one you thought that was off).

[phofl]

Not certain this is the PR that introduced the regression. Is it possible that making the shallow copy improves performance?

Yep possible, the shallow copy sets up references which means that follow up operations are not inplace anymore, which is sometimes faster. I would close here since this preserves the semantics that we want

[rhshadrach]

Thanks for checking.