Bug with iterating df.columns

[jonashaag]

Describe the bug

import pandas as pd

t: pd.DataFrame
s: list[str] = list(t.columns)

t.py:4: error: Argument 1 to "list" has incompatible type "Index"; expected
"Iterable[str]"
    s: list[str] = list(t.columns)
                        ^~~~~~~~~
t.py:4: note: Following member(s) of "Index" have conflicts:
t.py:4: note:     Expected:
t.py:4: note:         def __iter__(self) -> Iterator[str]
t.py:4: note:     Got:
t.py:4: note:         def __iter__(self) -> Iterator[Union[Union[str, bytes, date, datetime, timedelta, bool, int, float, complex, Timesta
mp, Timedelta], Tuple[Hashable, ...]]]
Found 1 error in 1 file (checked 1 source file)

Variant:

import pandas as pd

t: pd.DataFrame
s: str
for s in t.columns:
    pass

t.py:5: error: Incompatible types in assignment (expression has type
"Union[Union[str, bytes, date, datetime, timedelta, bool, int, float, complex, Timestamp, Timedelta], Tuple[Hashable, ...]]",
variable has type "str")
    for s in t.columns:
    ^
Found 1 error in 1 file (checked 1 source file)

To Reproduce
See above

Please complete the following information:

• Version 1.5.2.230105
• MyPy 0.991

[Dr-Irv]

You have to use cast here for now. We are not tracking the type of values inside of an Index. This is on our TODO list. But even with your example, there is no way to know the dtype of t.columns for any particular DataFrame.

You'd get a similar error in your first example like this:

import pandas as pd

t: pd.DataFrame
s: list[int] = list(t.columns)

So the solution here is to do:

import pandas as pd
from typing import cast

t: pd.DataFrame
s: list[str] = cast(list[str], list(t.columns))

[davetapley]

I also ran in to this using date_range.

Workaround is good for now:

for ts in cast(Iterator[datetime], date_range(start='1/1/2023', end='1/08/2023', freq='6H', tz=timezone.utc)):
   ....

[Dr-Irv]

I also ran in to this using date_range.

Different issue - this can be fixed. Created a new issue #723