Potential regression induced by PR #56921

[DeaMariaLeon]

PR #56921
Benchmarks:
join_merge.Join.time_join_dataframes_cross (Python) with sort=True
timeseries.ResetIndex.time_reset_datetimeindex (Python) with t=None
timeseries.ResetIndex.time_reset_datetimeindex (Python) with t='US/Eastern'
join_merge.Join.time_join_dataframes_cross (Python) with sort=False

@rhshadrach

[rhshadrach]

This is hitting the added check here:

pandas/pandas/core/internals/managers.py

Lines 1531 to 1534 in 737d390

	if (
	_get_option("performance_warnings")
	and sum(not block.is_extension for block in self.blocks) > 100
	):

It is hit twice for the entire benchmark; i.e. not in a tight loop. The regression is 0.2ms; this is expected. Closing.

[rhshadrach]

Reopening - this popped up on my regression tracker as well. However some of the times exceed 2ms, and should be investigated further.

• eval.Eval.time_chained_cmp
  • engine='numexpr'; threads='all'
• frame_methods.Shift.time_shift
  • axis=1
• indexing.InsertColumns.time_assign_list_like_with_setitem
• indexing.InsertColumns.time_assign_with_setitem
• indexing.InsertColumns.time_insert
• indexing.InsertColumns.time_insert_middle
• reshape.SparseIndex.time_unstack
• timeseries.ResetIndex.time_reset_datetimeindex
  • t='US/Eastern'
  • t=None