DEPR: groupby nuisance warnings

[rhshadrach]

Follow up to #41475

• All code checks passed.

The warning currently generated for pivot_table says groupby - but it points to the correct line of user code, and I think it's easy to understand what's going on from a user perspective. E.g.

df = pd.DataFrame(
    {
        "A": ["foo", "bar"],
        "B": ["small", "large"],
        "C": [1, 2],
        "D": ["3", "4"],
    }
)
result = pd.pivot_table(df, values=["C", "D"], index=["A"], columns=["B"], aggfunc="sum")

gives

scratch.py:61: FutureWarning: Dropping invalid columns in DataFrameGroupBy.sum is deprecated. In a future version, a TypeError will be raised. Before calling .sum, select only columns which should be valid for the function.
result = pd.pivot_table(df, values=["C", "D"], index=["A"], columns=["B"], aggfunc="sum")

cc @jbrockmendel

[jbrockmendel]

can include in previous line

[jbrockmendel]

comment here similar to L256, L259?

[rhshadrach]

The deprecations are essentially correct, but for the wrong reason. They should be deprecated due to #46072 and not #41475. Will be reworking this.