BUG: loc.__getitem__ stricter than it should be with list indexers

Describe the bug

loc.__getitem__ is stricter than it should be I think

error: List item 0 has incompatible type "int"; expected "str"  [list-item]
error: List item 0 has incompatible type "None"; expected "str"  [list-item]

To Reproduce

1. Provide a minimal runnable pandas example that is not properly checked by the stubs.
2. Indicate which type checker you are using (mypy or pyright).
3. Show the error message received from that type checker while checking your example.

df = pd.DataFrame([[1, 2]], columns=Index([1, None], dtype=object))

df.loc[:, [None]]
df.loc[:, [1]]

Alternatively you could define df as

df = pd.DataFrame({"a": [1, 2], None: 5})

Both of them should be valid. Especially since

df.loc[:, None]
df.loc[:, 1]

both work.

Please complete the following information:

• OS: ubuntu
• OS Version 22.04
• python version 3.10
• version of type checker mypy 0.971
• version of installed pandas-stubs newest 20220807

Additional context Add any other context about the problem here.

[Dr-Irv]

I think thats a bit different, the list elements are restricted to str, was this deliberate?

No, it’s just a result of how the types evolved over time by Microsoft.

I guess we have to change StrLike to Hashable or HashableT based on whether it is “alone” (use Hashable) or as a parameter to a type, e.g., list[StrLike] becomes list[HashableT]

[phofl]

It does, but you have to use different column names. You can either use a non integer as the other columns or do columns=Index([1, None], dtype=object)

if the column is a true None and not NaN this returns the column as expected

Edit: Sorry for the shit example in the op, don’t know how I mixed that up. Updated it now