RLS: 3.0

[datapythonista]

Estimated release date: April 2024, but it will be updated closer to that time.

[mroeschke]

Looking like this this release will happen post April 2024. Kicking the tires on things that are left to do before 3.0

Blockers for 3.0.

• Enforcing deprecations in 2.x DEPR: List of deprecations to be removed in 3.0 #50578
  • Except (potentially) resampling with PeriodIndex QST: FutureWarning: Resampling with a PeriodIndex is deprecated, how to resample now? #57033, QST: Roadmap for deprecations of Period types #56588
• Doing some action around pyarrow as dependency & the default string type DISC: Consider not requiring PyArrow in 3.0 #57073, FEEDBACK: PyArrow as a required dependency and PyArrow backed strings #54466

Notable non-blockers for 3.0

• Drop Python 3.9 support per NEP 29 and SPEC-0
• Separate out pandas tests into it's own package BLD: Split out tests into pandas_tests package #53007
• Deprecate make_block and introduce alternative DEPR: make_block #56815, Add low-level create_dataframe_from_blocks helper function #58197

Anything else @pandas-dev/pandas-core

[lithomas1]

It's probably wise to get in a 2.2.3 release before 3.0.

[rhshadrach]

Anything else pandas-dev/pandas-core

A minor issue, but I'd like some eyes on #57351, make sure we're still good with the removal of swapaxes.

[Aloqeely]

Anything else

If PDEP-17 goes through it'd be great if I could to switch all our deprecations from FutureWarning to DeprecationWarning before 3.0 — not a blocker though

[EwoutH]

Estimated release date: April 2024

Any update on the estimated release date? And will there be alpha, beta and/or release-candidate releases?

[BastienGauthier]

Hello everyone ! I hear everywhere that pandas 3.0 will be awesome, bravo !
Do you have a better view on when we should expect a release (or rc) ? Is it next month/quarter/year ?
Thanks again for the fantastic work done !

[jbrockmendel]

Can't speak for everybody, but I think we're getting close and am optimistic about an RC in August or September.