(fix): extension array indexers

[ilan-gold]

Identical to kmuehlbauer#1 - probably not very helpful in terms of changes since https://github.com/kmuehlbauer/xarray/tree/any-time-resolution-2 contains most of it....

• Closes #Failure in pandas TestDataFrameToXArray.test_to_xarray_index_types #9661
• Tests added
• User visible changes (including notable bug fixes) are documented in whats-new.rst
• New functions/methods are listed in api.rst

[benbovy]

Unfortunately I'm not able to review all changes related to datetime types, but I just wanted to note that the changes in core/indexes.py and core/indexing.py will likely solve xarray-contrib/xvec#87.

Any blocker apart from the CI failures and merge conflicts?

Doctest errors should be trivial to fix (although looking at the errors this makes me think that pandas Extension types might cause some rendering issues of Xarray variable inline reprs, depending on how long the reprs of the extension types can be).

[kmuehlbauer]

@benbovy Those changes have been added on top of #9618, which has diverged from the contents here. I'm not sure what the best procedure is, but maybe disentangling this PR from #9618 would work?

[ilan-gold]

@benbovy Those changes have been added on top of #9618, which has diverged from the contents here. I'm not sure what the best procedure is, but maybe disentangling this PR from #9618 would work?

I'm happy to update the PR here

[ilan-gold]

maybe disentangling this PR

The reason I created my PR on top of yours was that I spent many hours trying to do what you did, but much worse. The issue is that datetime + extension array permissibility are quite interconnected.

[ilan-gold]

@kmuehlbauer If you fix the merge conflicts with main I can clean up the diff in the PR to your repo so people can review there. And then we'll have CI here

[kmuehlbauer]

If you fix the merge conflicts with main I can clean up the diff in the PR to your repo so people can review there. And then we'll have CI here

Thanks @ilan-gold for the details. I've fixed the conflicts over in #9618. Would be great to get this over the line soon :-)

[ilan-gold]

@benbovy Cleaner diff over in https://github.com/kmuehlbauer/xarray/pull/1/files

[benbovy]

Thanks @ilan-gold. After a brief look it looks like kmuehlbauer#1 could be opened here as a separate PR targeting pydata/xarray/main, or maybe I'm wrong?

[ilan-gold]

@benbovy Is that not what this PR is?

[benbovy]

I was actually referring to #9671 (comment), but waiting for #9618 to be merged may be fine too I guess. The cleaner diff over in https://github.com/kmuehlbauer/xarray/pull/1/files makes the review easier although CI here still mixes those changes with #9618.

[ilan-gold]

CI here still mixes those changes with #9618.

Yeah, it's unfortunate but so many datetime indexers are passed through now and left untouched as a result of the changes here that I either had to restrict the datetime types one-by-one where they pop up or re implement the entirety of #9618 anyway to loosen the restriction globally.

[ilan-gold]

Everything working except small mypy stuff and a weird test or two that seem dependent on python version: 8a3e834

I will be on vacation for two-ish weeks starting tomorrow but this PR is probably reviewable, at least in hte other repo and I am happy to work on it as soon as I'm back.

[kmuehlbauer]

Everything working except small mypy stuff and a weird test or two that seem dependent on python version: 8a3e834

#9873