pandas.Index.to_numpy() method's arguments not properly typed

Describe the bug The pandas.Index.to_numpy() method takes arguments, but mypy complains about them, whether they are positional or keyword.

To Reproduce

import pandas as pd df = pd.DataFrame({“A”: [1, 2, 3], “B”: [4, 5, 6]}) array = df.index.to_numpy(dtype=“int32”)

This yields error: Unexpected keyword argument “dtype” for “to_numpy” of “Index”

import pandas as pd df = pd.DataFrame({“A”: [1, 2, 3], “B”: [4, 5, 6]}) array = df.index.to_numpy(dtype=“int32”)

This yields error: Too many arguments for “to_numpy” of "Index

Please complete the following information:

• OS: Windows
• OS Version 10
• python version 3.10
• version of type checker: latest
• version of installed pandas-stubs1.4.3.220829

[Dr-Irv]

closed via #270

[Dr-Irv]

With respect to the tests not having check(assert_type(...)), that’s a legacy issue. The original stubs didn’t have them, but when we are adding new tests, we are requiring it. Eventually, all the tests should use that paradigm. Would be good if you fix all the to_numpy() tests to use that pattern.

Correct me if I’m wrong please, I should add similar functions:

• for indexes, create an index, call to_numpy and assert the type
• for series, create a serie, call to_numpy and assert the type

Yes, you should add similar functions for all the to_numpy() variants.

Also. It seems like there is an addition here. Should I include the addition in the current PR or create a different one ?

You can do one PR, that includes fixes to all the to_numpy() stubs, and the appropriate tests.

Thanks in advance for your contribution!