Potential regression induced by PR #57588

[DeaMariaLeon]

PR #57588

If this is expected please ignore this issue.

The largest regressions seem to be the benchmarks bellow.
edit: the names of all the benchmarks with links to their plot are in the attached file.

series_methods.Dropna.time_dropna (Python) with dtype='datetime'
groupby.GroupByMethods.time_dtype_as_field (Python) with application='direct', dtype='int16', method='head', ncols=1, param5='cython'
groupby.Nth.time_series_nth_all (Python) with dtype='datetime'

@mroeschke

cleaned_regression_file.json

[mroeschke]

Thanks for the tag. This Is expected as there is more work done to potentially return a RangeIndex instead of an Index but I'll double check once more if some of that work can be minimized