No overload variant of "apply" of "DataFrameGroupBy" matches argument types "Callable[[Any], Any]", "bool"

[zyassine]

Describe the bug
Regression in the latest version

To Reproduce

df = pd.DataFrame({"category": ["A", "A", "B", "B"], "value": [10, 20, 30, 40]})
result = df.groupby("category", observed=True).apply(
        lambda x: pd.Series({"mean_value": x["value"].mean()}), include_groups=False
)

Using pandas 2.2.3, mypy 1.15.0 and pandas-stubs 2.3.0.250703

[twoertwein]

Thank you for reporting!

This change was somewhat on purpose #1268 as include_groups is "partly" deprecated. I think the idea is that this argument might no longer exist in the future. As long as True is the default value of include_groups , it would make sense to keep allowing include_groups=False.

[Dr-Irv]

df = pd.DataFrame({"category": ["A", "A", "B", "B"], "value": [10, 20, 30, 40]})
result = df.groupby("category", observed=True).apply(
lambda x: pd.Series({"mean_value": x["value"].mean()}), include_groups=False
)

The recommended way to get around this issue is to do the following:

df = pd.DataFrame({"category": ["A", "A", "B", "B"], "value": [10, 20, 30, 40]})
result = df.groupby("category", observed=True)[["value"]].apply(
        lambda x: pd.Series({"mean_value": x["value"].mean()})

or at least for the above computation:

 result = df.groupby("category", observed=True)[["value"]].mean().rename(columns={"value":"mean_value"})

would give the same result and probably be more performant.

So I'm going to close this