pandas-dev/pandas · error · TypeError

Accumulation not supported for

Error message

Accumulation {name} not supported for {type(self)}

What it means

Raised in Categorical._accumulate when the requested accumulation name is neither 'cummin' nor 'cummax'. Categoricals only support order-based accumulations; arithmetic accumulations like cumsum/cumprod are meaningless for non-numeric category labels, so they are rejected. The check fires before the ordered check, so even an ordered categorical will fail for cumsum.

Solutions

  1. For cummin/cummax, mark the categorical ordered (see error 240).
  2. For arithmetic accumulations, convert to the underlying numeric dtype first: df['col'].astype('int64').cumsum() — only valid if categories are numeric.
  3. Drop or exclude categorical columns before calling df.cumsum() across the frame.
  4. Use .cat.codes if you need integer accumulation over code positions (semantics differ).

Example fix

# before
s = pd.Series(pd.Categorical(['a','b','c'], ordered=True))
s.cumsum()  # TypeError: Accumulation cumsum not supported

# after (if categories are numeric)
s = pd.Series(pd.Categorical([1,2,3], ordered=True))
s.astype('int64').cumsum()
Defensive patterns

Strategy: type-guard

Validate before calling

name = 'cumsum'
if isinstance(s.dtype, pd.CategoricalDtype) and name not in ('cummin', 'cummax'):
    raise ValueError(f'{name} not supported on categorical; cast first')
s.__getattribute__(name)()

Type guard

def supports_accumulate(s: pd.Series, name: str) -> bool:
    if isinstance(s.dtype, pd.CategoricalDtype):
        return name in ('cummin', 'cummax')
    return True

Try / catch

try:
    s.cumsum()
except TypeError as e:
    if 'not supported for' in str(e) and isinstance(s.dtype, pd.CategoricalDtype):
        s.astype('int64').cumsum()
    else:
        raise

Prevention

When it happens

Trigger: df['catcol'].cumsum(), df['catcol'].cumprod(), df['catcol'].cummin() on an unordered categorical (this error fires only for names other than cummin/cummax), np.add.accumulate on a categorical. df.cummax() on a DataFrame containing a categorical column (the categorical's _accumulate is dispatched to).

Common situations: User calls df.cumsum() on a wide DataFrame that includes a string-typed categorical column. Migrating code that previously ran on object dtype where cumsum silently concatenated. Expecting numeric behavior from a categorically-typed column of integer categories (still rejected because accumulation is by category, not arithmetic).

Related errors

AI-assisted analysis of pandas-dev/pandas@3b7651241d (2026-08-11). Data as JSON: /api/errors/590bdea6829af19f. Report an issue: GitHub.

Appendix: source

Thrown at pandas/core/arrays/categorical.py:2671

        Returns
        -------
        bool
        """
        if not isinstance(other, Categorical):
            return False
        elif self._categories_match_up_to_permutation(other):
            other = self._encode_with_my_categories(other)
            return lib.array_equivalent_bytes(self._codes, other._codes)
        return False

    def _accumulate(self, name: str, skipna: bool = True, **kwargs) -> Self:
        func: Callable
        if name == "cummin":
            func = np.minimum.accumulate
        elif name == "cummax":
            func = np.maximum.accumulate
        else:
            raise TypeError(f"Accumulation {name} not supported for {type(self)}")
        self.check_for_ordered(name)

        codes = self.codes.copy()
        mask = self.isna()
        if func == np.minimum.accumulate:
            codes[mask] = np.iinfo(codes.dtype.type).max
        # no need to change codes for maximum because codes[mask] is already -1
        if not skipna:
            mask = np.maximum.accumulate(mask)

        codes = func(codes)
        codes[mask] = -1
        return self._simple_new(codes, dtype=self._dtype)

    @classmethod
    def _concat_same_type(cls, to_concat: Sequence[Self], axis: AxisInt = 0) -> Self:
        from pandas.core.dtypes.concat import union_categoricals

View on GitHub (pinned to 3b7651241d)