pandas-dev/pandas · error · NotImplementedError

No masked accumulation defined for dtype {values.dtype.type}

Error message

No masked accumulation defined for dtype {values.dtype.type}

What it means

Raised by _cum_func in masked_accumulations.py:52 as a NotImplementedError when a masked accumulation is requested on an array whose dtype kind is not f (float), i (signed int), u (unsigned int), or b (bool). The masked accumulation path computes min/max fill values from np.iinfo/np.finfo, which only exist for those kinds; any other dtype (object, str, complex, datetime) is unsupported and rejected up front.

Source

Thrown at pandas/core/array_algos/masked_accumulations.py:52

        Numpy array with the values (can be of any dtype that support the
        operation).
    mask : np.ndarray
        Boolean numpy array (True values indicate missing values).
    skipna : bool, default True
        Whether to skip NA.
    """
    dtype_info: np.iinfo | np.finfo
    if values.dtype.kind == "f":
        dtype_info = np.finfo(values.dtype.type)
    elif values.dtype.kind in "iu":
        dtype_info = np.iinfo(values.dtype.type)
    elif values.dtype.kind == "b":
        # Max value of bool is 1, but since we are setting into a boolean
        # array, 255 is fine as well. Min value has to be 0 when setting
        # into the boolean array.
        dtype_info = np.iinfo(np.uint8)
    else:
        raise NotImplementedError(
            f"No masked accumulation defined for dtype {values.dtype.type}"
        )
    try:
        fill_value = {
            np.cumprod: 1,
            np.maximum.accumulate: dtype_info.min,
            np.cumsum: 0,
            np.minimum.accumulate: dtype_info.max,
        }[func]
    except KeyError as err:
        raise NotImplementedError(
            f"No accumulation for {func} implemented on BaseMaskedArray"
        ) from err

    values[mask] = fill_value

    if not skipna:
        mask = np.maximum.accumulate(mask)

View on GitHub (pinned to 71959b8cb9)

Solutions

1. Cast the data to a supported numeric dtype before accumulating: s.astype('Int64').cumsum() or s.astype('Float64').cumsum().
2. Use the appropriate accumulation for the dtype - datetimelike arrays have their own accumulator (datetimelike_accumulations).
3. For EA authors: implement cumsum/cummin/etc. directly on your array subclass instead of routing non-numeric dtypes through the shared masked path.

Example fix

// before
arr = pd.arrays.IntegerArray(np.array(['1','2'], dtype=object), np.array([False,False]))
arr.cumsum()  # unsupported kind
// after
s = pd.Series(['1','2']).astype('Int64')
s.cumsum()

kind = getattr(values, 'dtype', type(values)).kind if hasattr(values, 'dtype') else None
if kind not in 'fiub':
    raise NotImplementedError(f'masked accumulation unsupported for dtype kind {kind!r}; cast to numeric first')

def masked_accum_dtype_ok(values) -> bool:
    import numpy as np
    dt = getattr(values, 'dtype', None)
    return dt is not None and dt.kind in 'fiub'

When it happens

Trigger: Calling cumsum/cumprod/cummin/cummax on a masked ExtensionArray (e.g. IntegerArray, FloatingArray, BooleanArray) backed by an unsupported dtype, or reaching the masked path with an object/complex array. Hit at masked_accumulations.py:42-54 when values.dtype.kind is none of f/i/u/b.

Common situations: Custom ExtensionArrays with non-numeric backing dtypes routed through the masked accumulation; converting an object-dtype column to a masked array and calling cumsum; bugs in EA dispatch that send datetime/string arrays into the numeric masked path.

Related errors

• No accumulation for {func} implemented on BaseMaskedArray
• {type(arr).__name__} has no 'diff' method. Convert to a suit
• Column {colname} must have a numeric dtype. Found '{dtype}'
• Column {colname} is backed by an extension array, which is n
• No accumulation for {func} implemented on BaseMaskedArray

AI-assisted analysis of pandas-dev/pandas@71959b8cb9 (2026-08-07). Data as JSON: /api/errors/95a1b41e213ee062. Report an issue: GitHub.