xref #30228

When opening a PR that introduces a deprecation (starts warning), please add it to this list, so we know what needs to be fully deprecated when releasing pandas 3.0.

2.0

• [x] #48454
• [x] #48735
• [x] #50176
• [x] #50178
• [x] #50196
• [x] #50235
• [x] #50769
• [x] #50227
• [x] #50225
• [x] #50243
• [x] #50483
• [x] #50676
• [x] #51182
• [x] #51042
• [x] #51019
• [x] #51221
• [x] #51206
• [x] #52071
• [x] #50947

2.1

• [x] #51539
• [x] #51477
• [x] #51697
• [x] #51395
• [x] #51703
• [x] #51756
• [x] #51811
• [x] #51968
• [x] #52003
• [x] #52257
• [x] #52268
• [x] #52333
• [x] #52364
• [x] #52459 (change behavior)
• [x] #53419
• [x] #53325
• [x] #53493
• [x] #53607
• [x] #53638
• [x] #53492 (closed by #58039)
• [x] #53571 (closed by #57136)
• [x] #53557
• [x] #53409
• [x] #53710
• [x] #53830
• [x] #53805
• [x] #53671
• [x] #53994
• [x] #53975
• [x] #53956
• [x] #53974
• [x] #52532
• [x] #53817
• [x] #53809
• [x] #52788 (remove keepdims check inside blk_func in DataFrame._reduce)
• [x] #54170
• [ ] #53511
• [x] #54226
• [x] #54219
• [x] #54014
• [x] #53338
• [x] #54397
• [x] #54599
• [x] #54601
• [x] #54597
• [x] #54613
• [x] #54605
• [x] #54155
• [x] #54612
• [x] #54632
• [x] #54631
• [x] #54663
• [x] #54634
• [x] #54630
• [x] #54704
• [x] #54703
• [x] #54818
• [x] #54061
• [x] #53656
• [x] #54950
• [x] #54710
• [x] #54261
• [x] #55044
• [x] #55313
• [x] #55252
• [ ] #54939
• [x] #55466
• [x] #55479
• [x] #55439
• [x] #55496
• [x] #55448
• [ ] #55355
• [ ] #52419
• [x] #55636
• [x] #55553
• [x] #55734
• [x] #53769
• [x] #55792
• [x] #55868
• [x] #55940
• [x] #55963
• [x] #55719
• [x] #55978
• [x] #56038
• [x] #56050
• [x] #56055
• [ ] #55139
• [x] #56036
• [x] #56092
• [x] #56161 (closed by #58009)
• [x] #56149
• [x] #56054
• [x] #56053
• [x] #55895
• [x] #55856 (closed by #58003)
• [x] #56237
• [ ] #55703
• [x] #55257
• [x] #56385
• [x] #56427
• [x] #56421 (closed by #58047)
• [x] #56102
• [x] #56489
• [ ] #55968
• [x] #56521
• [x] #56556
• [x] #56558 (closed by #58005)
• [x] #56569
• [x] #56557
• [x] #56043 (closed by #58004)
• [x] #56244
• [x] #56594
• [x] #53496

2.3

• [ ] #59615

Additional: - [ ] #59240 - [ ] https://github.com/pandas-dev/pandas/pull/59051 - checked off above, but I'm seeing remnants in timedeltas.pyx. - [x] #61264 - [ ] #60152 - [ ] #60234 - [ ] #59561

Comment From: topper-123

50052 should not be om this list, it is part of the pandas 1.x -> 2.0 cleanup process.

Comment From: topper-123

Likewise #49560.

EDIT: ive removed both of them from the list.

Comment From: DeaMariaLeon

Could you add this one to the list please?: DEPR: Series.first() and DataFrame.first()

53419

@datapythonista

Comment From: topper-123

I've added it, thanks @DeaMariaLeon.

More generally, I can see several deprecatons in the pulse that have not been added, e.g. #52986 & #52981. To my knowledge all deprecations should be added to the list, or has that changed?

Comment From: topper-123

@pandas-dev/pandas-core, looks like we need to be more mindful of adding deprecations to this list. Or is there an alternative method of knowing which deprecations have been done?

Comment From: mroeschke

There was supposed to be a job that @phofl created that would update the OP https://github.com/pandas-dev/pandas/blob/main/.github/workflows/deprecation-tracking-bot.yml

Has it not been working recently?

Comment From: lithomas1

I don't think it ever worked except on his repo for some reason. Looks like there was a permissions issue somewhere.

Here's the log of a recent run https://github.com/pandas-dev/pandas/actions/runs/5195752658/jobs/9368776352

Comment From: lithomas1

Ok, here's my best guess to what's going wrong.

So Github typically restricts the types of perms a workflow can get if it comes from a fork (you get downgraded to read access). https://docs.github.com/en/actions/using-jobs/assigning-permissions-to-jobs#:~:text=permissions%3A%20%7B%7D-,You%20can%20use%20the%20permissions%20key%20to%20add%20and%20remove%20read,more%20information%2C%20see%20%22Managing%20GitHub%20Actions%20settings%20for%20a%20repository.%22,-Example%3A%20Assigning%20permissions

When Patrick tested his code on his fork, he opened a PR from a branch within the fork, so the PR was able to get write perms since it came from the same repo. However, when we PR to the main pandas repo, we do it from our forks, not from branches in the main repo, triggering the permissions downgrade.

I think the solution is to run the workflow with a push trigger, not a pull_request trigger, since only pull_request triggers the permission downgrade.

Probably need https://docs.github.com/en/rest/commits/commits?apiVersion=2022-11-28#list-pull-requests-associated-with-a-commit to map the commit back to the original PR.

I'll try to have something up soonish.

Comment From: lithomas1

It works 🎉 ! (Bot added #53493)

One thing to keep in mind though is that adding the "Deprecate" label is necessary to trigger the bot. (It doesn't trigger when the commit message/PR title has "DEPR" in it).

I've added #53607 manually, hopefully there's not anything else missing.

Comment From: phofl

Thx for fixing the bot!

Comment From: jbrockmendel

@natmokval can you check off the boxes above that you have already handled. i think you've done a bunch of them

Comment From: natmokval

@natmokval can you check off the boxes above that you have already handled. i think you've done a bunch of them

thanks @jbrockmendel. I tried to check the boxes off, but it didn't work. Maybe I don't have permissions to do it.

Comment From: rhshadrach

@natmokval - that's surprising, I would think triage permissions would allow you to edit any comment. Are you able to edit this comment?

In the meantime, you could list the PRs to be checked off in a comment here and someone else can update the OP.

Comment From: jbrockmendel

darn, thanks for taking a look.

Comment From: natmokval

@natmokval - that's surprising, I would think triage permissions would allow you to edit any comment. Are you able to edit this comment?

no, I can't edit this comment. I am allowed to edit only my comments.

In the meantime, you could list the PRs to be checked off in a comment here and someone else can update the OP.

thanks, @rhshadrach. Here the list of PR’s to be checked off:

• [x] #53638
• [x] #55479
• [x] #55252
• [x] #53557
• [x] #54014

I noticed a couple of missing PRs, which are need to be checked off too: - [x] #50947 - [x] #54061

the follow PR is also not on the list: - [ ] #53496

but I am not finished with the removal yet.

Comment From: rhshadrach

Thanks @natmokval - I've checked off all PRs.

I noticed a couple of missing PRs, which are need to be checked off too: - [x] #50947 - [x] #54061

The 2nd I found in the OP list, however I did not find the first. Should it be added?

Comment From: natmokval

Thanks @natmokval - I've checked off all PRs.

thanks @rhshadrach.

I noticed a couple of missing PRs, which are need to be checked off too:

• [x] DEPR: dt64 any/all GH#34479 #50947
• [x] DEPR: deprecate strings T, S, L, U, and N in offsets frequencies, resolution abbreviations, _attrname_to_abbrevs #54061

The 2nd I found in the OP list, however I did not find the first. Should it be added?

yes, I think we should add the first PR in the OP list in 2.0 section. I enforced this deprecation in #58029.

Comment From: rhshadrach

Thanks @natmokval - done.

Comment From: Aloqeely

@rhshadrach there are a few more that need to be checked off (correct me if I'm wrong):

• [ ] DEPR: to_pydatetime return Index[object] #52459 (change behavior)

enforced by #57425

• [ ] DEPR: concat ignoring empty objects #52532
• [ ] DEPR: Deprecate use of un-supported numpy dt64/td64 dtype for pandas.array #53817

by #58314 and #58060 respectively

• [ ] DEPR: argmin/argmax with all-NA or skipa=False #54170

by #57971

• [ ] DEPR: idxmin/idxmax with all-NA #54226

also by #57971

• [ ] DEPR: BaseNoReduceTests #54663

by #58053

• [ ] DEP: Deprecate passing fill_value and freq to shift #54818

by #57425

• [ ] TYP: overload for DataFrame.to_xml #55313

I can't find where, but I think this has been enforced

• [ ] DEPR: fractional periods in date_range, timedelta_range, period_range… #56036

by #57999

• [ ] DEPR: kind keyword in resample #55895

by #58125

• [ ] DEPR: Index.insert dtype-inference #55257

by #58059

• [ ] DEPR: make_block #56422

Should be removed from the list, was reverted in #56481

• [ ] DEPR: verbose kwd in read_csv, read_table #56556

by #57966

Comment From: rhshadrach

Thanks @Aloqeely! Updated the OP.

I also removed https://github.com/pandas-dev/pandas/pull/52477 from the list. That needs to be made a FutureWarning before we can enforce the deprecation.

Comment From: natmokval

Here is the list of PR’s enforced by CLN: enforce deprecation of frequencies deprecated for offsets #57986:

• [x] #56092
• [x] #55978
• [x] #56050
• [x] #55496
• [x] #55792
• [x] #55553

@rhshadrach, could you please check them off?

• [ ] #56910

should not be on this list, because it is not a deprecation. This PR is s a correction ValueError to FutureWarning. Could you please remove this PR from the list?

Comment From: rhshadrach

Thanks @DeaMariaLeon - done.

Comment From: rhshadrach

I believe the OP is now up to date. I added deprecations that were in the code but not the list and checked off ones that were enforced. Still a few to carry out here, cc @pandas-dev/pandas-core if people have any bandwidth.