BUG: `columns` on existing `DataFrame` does not accept "List[str]"

[bashtage]

import pandas as pd
df = pd.DataFrame({"a":[1,2,3],"b":[0.0,1,1]})

df.columns= ["c", "d"]  # mypy error, columns is `Index` so cannot take List[str}

[Dr-Irv]

This is a bug in mypy: python/mypy#3004

Here's a simple example:

from typing import Union


class Foobar:
    def __init__(self, x: Union[int, str]):
        self._x = self.makestr(x)

    def makestr(self, x: Union[int, str]) -> str:
        if isinstance(x, int):
            return str(x)
        elif isinstance(x, str):
            return x
        else:
            raise Exception("invalid type")

    @property
    def x(self) -> str:
        return self._x

    @x.setter
    def x(self, x: Union[int, str]) -> None:
        self._x = self.makestr(x)


fb = Foobar("abc")
fb.x = 3

mypy will then report:

mypyasym.py:26: error: Incompatible types in assignment (expression has type "int", variable has type "str")

[Dr-Irv]

mypy has fixed this bug in python/mypy#18510, so once that is released we can remove the # type: ignore comments in test_frame.py:test_set_columns(). Seems to be OK in this version /mypy-1.16.0%2Bdev.42e005c999d8341c0da6d7b93b10d05f2db2099 so we'll have to keep an eye out for the version 1.16 release.

[Dr-Irv]

closed via #1236