[index] DataFrame.loc does not accept tuple[slice, tuple]

[randolf-scholz]

Describe the bug

The current type hint contains:

tuple[
    IndexType | MaskType | list[HashableT] | Hashable,
    list[HashableT] | slice | Series[bool] | Callable
]

The following modification should be considered

1. Ideally, replace list[HashableT] with Sequece[HashableT]. But this doesn't work because str satsifies Sequence protocol. So at the least one could consider adding tuple manually.
   • Since this issue seems to be common, one might consider creating a custom container TypeAlias that in the future could be replaced with Sequence[H] & ~H if we ever get intersection and not types.
2. add slice to the first component.

Suggestion

tuple[
    # in the index `Iterable[H]` works fine!
    slice | Hashable | Iterable[HashableT] | IndexType | MaskType,
    # For the columns use `list[HashableT] | tuple[HashableT, ...]` instead of `Iterable[H]` to avoid matching `str`.
    slice | list[HashableT] | tuple[HashableT, ...] | Series[bool] | Callable,  
]

To Reproduce

import numpy as np
from pandas import DataFrame

foo = DataFrame(np.random.rand(10,3), columns=["a", "b", "c"])
bar = foo.loc[4:5, ("a", "b")]  # [index]

Please complete the following information:

• pandas 1.5.3
• pandas-stubs 1.5.3.230304

[randolf-scholz]

Actually, if one restricts to loc[slice, list] it works, issue is with Multiindex columns allowing tuple to index a single column, which intersects with selecting multiple columns by an iterable of hashables. Closing.