Align typing of dimension inputs

[headtr1ck]

What is your issue?

Currently the input type for "one or more dims" is changing from function to function. There are some open PRs that move to str | Iterable[Hashable] which allows the use of tuples as dimensions.

Some changes are still required:

• [ ] Accept None in all functions that accept dims as default, this would simplify typing alot (see https://github.com/pydata/xarray/pull/7048#discussion_r973813607)
• [ ] Check if we can always include ellipsis "..." in dim arguments (see https://github.com/pydata/xarray/pull/7048#pullrequestreview-1111498309)
• [ ] Iterable[Hashable] includes sets, which do not preserve the ordering (see https://github.com/pydata/xarray/pull/6971#discussion_r981166670). This means we need to distinguish between the cases where the order matters (constructor, transpose etc.) and where it does not (drop_dims, reductions etc.). Probably this needs to be typed as a str | Sequence[Hashable] (a numpy.ndarray is not a Sequence, but who uses this for dimensions anyway?).

[derhintze]

Iterables (implement __iter__) not only aren't ordered, but also neither have a length (implement __len__ nor allow for membership tests (implement __contains__). So in these cases one needs to use Sequences, too; in addition to cases where order matters.

See https://docs.python.org/3.9/library/collections.abc.html#collections-abstract-base-classes

[headtr1ck]

Yes, sequence sounds right if the order matters. Usually we do not use len(input) or x in input though. Maybe reversible container or something else exists?

A bit unfortunate that np.ndarray is not a sequence though, but I think using them for dimensions of not a very common use case.

[max-sixty]

Usually we do not use len(input) or x in input though. Maybe reversible container or something else exists?

We can build a list if we need a len? Or IIRC there is a Sized type — a Set but not an Iterable.

[headtr1ck]

We can build a list if we need a len? Or IIRC there is a Sized type — a Set but not an Iterable.

We already do that.

We have the following solutions:

1. We allow any Iterable and it is up to the user to ensure correct ordering (mypy does not complain when a set is supplied)
2. We require a Sequence which ensures ordering (user cannot supply a np.ndarray e.g.)
3. We make the type more complicated to support ArrayLike as well.

[max-sixty]

Ah excellent, thanks @headtr1ck