type multiindex constructors

[MarcoGorelli]

• Closes #xxxx (Replace xxxx with the Github issue number)
• Tests added: Please use assert_type() to assert the type of any return value

[MarcoGorelli]

This one's a bit odd

The name is from_arrays, and the pandas docstring says that this should be a sequence of array-like

But, the pandas docs (and tests) have several examples of passing lists, which are not ArrayLike

lists are not included in AnyArrayLike, so I've used SequenceNotStr[Hashable] | AnyArrayLike

[MarcoGorelli]

🤔 maybe ListLike would be more suitable

[Dr-Irv]

I think it should be Sequence[Axes] because this works:

>>> pd.MultiIndex.from_arrays([range(2), [3,4]], names=["a", "b"])
MultiIndex([(0, 3),
            (1, 4)],
           names=['a', 'b'])
>>> pd.MultiIndex.from_arrays([{1:"x", 2:"y"}, [3,4]], names=["a", "b"])
MultiIndex([(1, 3),
            (2, 4)],
           names=['a', 'b'])

[Dr-Irv]

🤔 maybe ListLike would be more suitable

Problem is that ListLike allows a single string, so we can't use it. I think Axes is the right choice.

[Dr-Irv]

I think it should be Sequence[Axes] because this works:

>>> pd.MultiIndex.from_arrays([range(2), [3,4]], names=["a", "b"])
MultiIndex([(0, 3),
            (1, 4)],
           names=['a', 'b'])
>>> pd.MultiIndex.from_arrays([{1:"x", 2:"y"}, [3,4]], names=["a", "b"])
MultiIndex([(1, 3),
            (2, 4)],
           names=['a', 'b'])

[MarcoGorelli]

thanks for your review and explanations!

[Dr-Irv]

thanks @MarcoGorelli