pandas-dev/pandas · error · TypeError

{dtype} type does not support {how} operations

Error message

{dtype} type does not support {how} operations

What it means

Raised by Categorical._groupby_op when the groupby aggregation 'how' is one of sum, prod, cumsum, cumprod, skew, or kurt. Arithmetic aggregations are undefined on category labels (they have no magnitude), so pandas rejects them up front before any computation. This is a hard type contract: categorical supports order and boolean aggregations only.

Source

Thrown at pandas/core/arrays/categorical.py:2871

    def _groupby_op(
        self,
        *,
        how: str,
        has_dropped_na: bool,
        min_count: int,
        ngroups: int,
        ids: npt.NDArray[np.intp],
        **kwargs,
    ):
        from pandas.core.groupby.ops import WrappedCythonOp

        kind = WrappedCythonOp.get_kind_from_how(how)
        op = WrappedCythonOp(how=how, kind=kind, has_dropped_na=has_dropped_na)

        dtype = self.dtype
        if how in ["sum", "prod", "cumsum", "cumprod", "skew", "kurt"]:
            raise TypeError(f"{dtype} type does not support {how} operations")
        if how in ["min", "max", "rank", "idxmin", "idxmax"] and not dtype.ordered:
            # raise TypeError instead of NotImplementedError to ensure we
            #  don't go down a group-by-group path, since in the empty-groups
            #  case that would fail to raise
            raise TypeError(f"Cannot perform {how} with non-ordered Categorical")
        if how not in [
            "rank",
            "any",
            "all",
            "first",
            "last",
            "min",
            "max",
            "idxmin",
            "idxmax",
        ]:
            if kind == "transform":
                raise TypeError(f"{dtype} type does not support {how} operations")

View on GitHub (pinned to 71959b8cb9)

Solutions

  1. Cast the column to its numeric value dtype before aggregating: df['cat_col'].astype(df['cat_col'].cat.categories.dtype).
  2. Drop arithmetic aggregations from the agg spec for categorical columns.
  3. Confirm the column should be categorical; if not, leave it as numeric and skip dtype='category'.

Example fix

// before
df.groupby('g')['cat'].sum()  # TypeError: category type does not support sum

// after
df.groupby('g')['cat'].astype('int64').sum()  # if categories are numeric
# or use a valid aggregation
df.groupby('g')['cat'].first()
Defensive patterns

Strategy: type-guard

Validate before calling

ARITH_HOW = {'sum','prod','cumsum','cumprod','skew','kurt'}
def safe_groupby_agg(s, how):
    import pandas as pd
    if isinstance(s.dtype, pd.CategoricalDtype) and how in ARITH_HOW:
        raise TypeError(f'{how} undefined on Categorical; cast to numeric first')
    return getattr(s, how)()

Type guard

import pandas as pd
from typing import Any

def supports_groupby_how(obj: Any, how: str) -> bool:
    if isinstance(getattr(obj, 'dtype', None), pd.CategoricalDtype):
        return how not in {'sum','prod','cumsum','cumprod','skew','kurt'}
    return True

Try / catch

try:
    df.groupby('g')['cat'].sum()
except TypeError as e:
    if 'does not support' in str(e) and 'operations' in str(e):
        df['cat'].astype('int64').groupby(df['g']).sum()
    else:
        raise

Prevention

When it happens

Trigger: df.groupby('g')['cat_col'].sum(), .prod(), .cumsum(), .cumprod(), .skew(), or .kurt() on a categorical-typed column.

Common situations: Blindly running df.groupby(...).agg(['sum','mean','min','max']) across all columns including object-like category columns; or a column that should be numeric but was inferred as category because it had few unique values.

Related errors

AI-assisted analysis of pandas-dev/pandas@71959b8cb9 (2026-08-07). Data as JSON: /api/errors/5db1919582a866cd. Report an issue: GitHub.