DataFrame.loc does not like assigning Enums

[cmp0xff]

Describe the bug

DataFrame.loc does not like assigning Enums.

I think this is related to #1288. I do not mean to be annoying, and the example came from our real use cases.

To Reproduce

Provide a minimal runnable pandas example that is not properly checked by the stubs.

from enum import Enum, auto

import pandas as pd


class Thing(Enum):
    A = auto()


data = pd.DataFrame(index=[0, 1])
data.loc[:, "a"] = Thing.A

Indicate which type checker you are using (mypy or pyright)

pyright

Show the error message received from that type checker while checking your example

attempt.py
  attempt.py:11:1 - error: No overloads for "__setitem__" match the provided arguments (reportCallIssue)
  attempt.py:11:1 - error: Argument of type "Literal[Thing.A]" cannot be assigned to parameter "value" of type "Scalar | ArrayLike | NAType | NaTType | Series[Any] | DataFrame | list[Unknown] | Mapping[Hashable, Scalar | NAType | NaTType] | None" in function "__setitem__"
    Type "Literal[Thing.A]" is not assignable to type "Scalar | ArrayLike | NAType | NaTType | Series[Any] | DataFrame | list[Unknown] | Mapping[Hashable, Scalar | NAType | NaTType] | None"
      "Literal[Thing.A]" is not assignable to "str"
      "Literal[Thing.A]" is not assignable to "bytes"
      "Literal[Thing.A]" is not assignable to "date"
      "Literal[Thing.A]" is not assignable to "datetime"
      "Literal[Thing.A]" is not assignable to "timedelta"
      "Literal[Thing.A]" is not assignable to "datetime64[date | int | None]"
      "Literal[Thing.A]" is not assignable to "timedelta64[timedelta | int | None]" (reportArgumentType)
2 errors, 0 warnings, 0 informations

Please complete the following information

• OS: Windows
• OS Version: 11
• python version: 3.13.4
• version of type checker: 1.1.404
• version of installed pandas-stubs: 2.3.2.250827

Additional context

None

[Dr-Irv]

It's very related. Only possible solution is to allow Enum as a valid Scalar. Would require a lot of tests, and also the arithmetic stuff should be handled correctly, because you can't do things like this:

from enum import Enum, auto
class Thing(Enum):
    A = auto()
    B = auto()
s1 = pd.Series([Thing.A, Thing.B])
s1 + s1

The addition fails.

So I think forcing a cast here on the user is the right thing to do, but if someone wants to do all the support for Enum inside of a Series, I'm fine with that.

[cmp0xff]

Thank you for the reply.

So I think forcing a cast here on the user is the right thing to do

You mean something like data.loc[:, "a"] = cast(ValidScalar, Thing.A)?

[Dr-Irv]

You mean something like data.loc[:, "a"] = cast(ValidScalar, Thing.A)?

This would work:

from pandas._typing import Scalar
data.loc[:, "a"] = cast(Scalar, Thing.A)