apache/beam · error · ValueError

Unsupported type for value of display data

Error message

Unsupported type %s for value of display data %s

What it means

Display data values must map to a protobuf LabelledPayload field (str, int, float, bool, timestamp, etc.). When converting DisplayData to its runner API proto, create_payload raises ValueError if the value's Python type has no supported payload representation.

Solutions

1. Return a supported primitive (str, int, float, bool) from display_data() for that item.
2. Convert the object with str(value) or a repr before returning it.
3. If it's a Beam bug, wrap the value or upgrade to a version where the type is supported.

Example fix

// before
def display_data(self):
    return {'conf': self.my_config_object}
// after
def display_data(self):
    return {'conf': str(self.my_config_object)}

SUPPORTED = (str, int, float, bool)
for k, v in self.display_data().items():
    assert isinstance(v, SUPPORTED), f'display data {k} has unsupported type {type(v)}'

def is_display_data_value(v) -> bool:
    return isinstance(v, (str, int, float, bool))

try:
    proto = display_data.to_proto()
except ValueError as e:
    if 'Unsupported type' in str(e):
        coerce_display_data_items_to_str(display_data)
        proto = display_data.to_proto()
    else:
        raise

When it happens

Trigger: A transform's display_data() returns an item whose value is an arbitrary object (e.g. a custom class instance, dict, or callable) that DisplayDataItem does not classify into a supported type, during to_proto() serialization.

Common situations: Adding display_data entries returning custom config objects; refactor leaving a non-primitive in a display item; library upgrades adding display data with new value types.

Related errors

• Attempted to encode null for non-nullable field
• Decode not implemented
• Element of class . does not subclass HasDisplayData
• Element of class . does not subclass PipelineOptions
• Encode not implemented

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/482bd8e31d1a7d93. Report an issue: GitHub.

Appendix: source

Thrown at sdks/python/apache_beam/transforms/display.py:180

        return beam_runner_api_pb2.LabelledPayload(
            label=label,
            bool_value=value,
            key=display_data_dict['key'],
            namespace=display_data_dict.get('namespace', ''))
      elif isinstance(value, int):
        return beam_runner_api_pb2.LabelledPayload(
            label=label,
            int_value=value,
            key=display_data_dict['key'],
            namespace=display_data_dict.get('namespace', ''))
      elif isinstance(value, (float, complex)):
        return beam_runner_api_pb2.LabelledPayload(
            label=label,
            double_value=value,  # type: ignore[arg-type]
            key=display_data_dict['key'],
            namespace=display_data_dict.get('namespace', ''))
      else:
        raise ValueError(
            'Unsupported type %s for value of display data %s' %
            (type(value), label))

    dd_protos = []
    for dd in self.items:
      if isinstance(dd, beam_runner_api_pb2.DisplayData):
        dd_protos.append(dd)
      else:
        dd_payload = create_payload(dd)
        if dd_payload:
          dd_protos.append(
              beam_runner_api_pb2.DisplayData(
                  urn=common_urns.StandardDisplayData.DisplayData.LABELLED.urn,
                  payload=dd_payload.SerializeToString()))
    return dd_protos

  @classmethod
  def create_from_options(cls, pipeline_options):

View on GitHub (pinned to 12126d8942)