apache/beam · error · TypeError

Element type is not a dtype

Error message

Element type is not a dtype

What it means

NumpyBatchConverter.from_typehints requires the element_type to be representable as a numpy dtype. If wrapping element_type as NumpyArray[element_type, ()] raises TypeError, the element type is not a valid numpy dtype and this error is raised.

Source

Thrown at sdks/python/apache_beam/typehints/batch.py:192

      batch_type,
      element_type,
      dtype,
      element_shape=(),
      partition_dimension=0):
    super().__init__(batch_type, element_type)
    self.dtype = np.dtype(dtype)
    self.element_shape = element_shape
    self.partition_dimension = partition_dimension

  @staticmethod
  @BatchConverter.register(name="numpy")
  def from_typehints(element_type,
                     batch_type) -> Optional['NumpyBatchConverter']:
    if not isinstance(element_type, NumpyTypeHint.NumpyTypeConstraint):
      try:
        element_type = NumpyArray[element_type, ()]
      except TypeError as e:
        raise TypeError("Element type is not a dtype") from e

    if not isinstance(batch_type, NumpyTypeHint.NumpyTypeConstraint):
      if not batch_type == np.ndarray:
        raise TypeError(
            "batch type must be np.ndarray or "
            "beam.typehints.batch.NumpyArray[..]")
      batch_type = NumpyArray[element_type.dtype, (N, )]

    if not batch_type.dtype == element_type.dtype:
      raise TypeError(
          "batch type and element type must have equivalent dtypes "
          f"(batch={batch_type.dtype}, element={element_type.dtype})")

    computed_element_shape = list(batch_type.shape)
    partition_dimension = computed_element_shape.index(N)
    computed_element_shape.pop(partition_dimension)
    if not tuple(computed_element_shape) == element_type.shape:
      raise TypeError(

View on GitHub (pinned to 12126d8942)

Solutions

1. Use a numpy-compatible scalar element type such as np.int64, np.float64, or str
2. Convert your element type to a registered numpy dtype before calling
3. Use a different BatchConverter (e.g. list) for non-dtype element types

Example fix

// before
BatchConverter.from_typehints(Union[int, str], np.ndarray)
// after
BatchConverter.from_typehints(np.int64, np.ndarray)

import numpy as np
try:
    np.dtype(element_type)
except TypeError:
    raise ValueError(f'{element_type!r} is not a numpy dtype')

def is_np_dtype(t):
    try:
        np.dtype(t); return True
    except TypeError:
        return False

try:
    conv = BatchConverter.from_typehints(elem_t, np.ndarray)
except TypeError:
    conv = BatchConverter.from_typehints(elem_t, List[elem_t])  # fallback to list converter

When it happens

Trigger: from_typehints with an element type like Union[int, str], Dict[str, int], or a custom Python class that numpy cannot interpret as a dtype.

Common situations: Passing beam or typing hints (e.g. Optional[int], Any) as element types to numpy batching; using structured types numpy doesn't accept.

Related errors

• batch type and element type must have equivalent dtypes (bat
• Batch {batch!r} does not have expected dtype: {self.dtype!r}
• batch type must be np.ndarray or beam.typehints.batch.NumpyA
• Failed to align batch type's batch dimension with element ty
• Batch {batch!r} is not an instance of ndarray

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/c3a8c1db02d09cb1. Report an issue: GitHub.