read_excel/index_col: string not accepted

[clo-vis]

Either the documentation or the type hint for index_col in pandas.read_excel is wrong.

The documentation says a string may be used.

The type hint forbids that

Documentation:
index_col int, str, list of int, default None

Type hints:
index_col: int | Sequence[int] | None = ...,

[Dr-Irv]

Looking internally at the pandas source, I think the docs may be wrong. Inside of pandas, index_col is typed without allowing str as an arg type.

If you can produce an example where read_excel() accepts a string (and maybe a list of strings??) as an argument for index_col, then updates should be made to the internal pandas source and the type hints.

Conversely, if a string doesn't work as the argument, then an issue should be raised in the pandas repo.

Can you provide an appropriate test (either way)?

[clo-vis]

from pandas import read_excel
print(read_excel("Book1.xlsx"))
print(read_excel("Book1.xlsx", index_col="bar"))

Output:

   foo  bar
0    1    2
1    3    4
     foo
bar     
2      1
4      3

Book1.xlsx

[Dr-Irv]

Thanks for the example. I created an issue for the pandas source to fix the typing there.

In the meantime, a fix here would be welcome, along with appropriate tests.