apache/beam · error · ValueError

Unrecognized type_info

Error message

Unrecognized type_info: {type_info!r}

What it means

`typing_from_runner_api` dispatches on the FieldType proto's WhichOneof('type_info'). If the discriminator is none of the known branches (atomic_type, array_type, map_type, row_type, iterable_type, logical_type, etc.), it raises this ValueError — the proto contains a type_info value this SDK version doesn't understand.

Solutions

1. Upgrade apache-beam to match the producer's protocol version
2. Identify the unexpected type_info value from the message and restructure the schema to use supported types
3. Regenerate/re-export the schema with an older-compatible SDK

info = fieldtype_proto.WhichOneof('type_info')
known = {'atomic_type','array_type','map_type','row_type','iterable_type','logical_type'}
assert info in known, f'unknown type_info {info}'

try:
  pytype = converter.typing_from_runner_api(fieldtype)
except ValueError as e:
  raise SchemaDecodeError(str(e)) from e

When it happens

Trigger: Decoding a FieldType proto with an unrecognized type_info, typically protos emitted by a newer Beam runner/SDK or another language SDK using a type not yet supported in Python.

Common situations: Mixed SDK versions in cross-language pipelines; deserializing cached schema protos written by a newer Beam release.

Understand the failure class

Background: Invalid enum value errors: "Unknown type", "Invalid scope", "must be one of" — when a string is not on the library's allowed list — this error's family across 23 libraries.

Related errors

• Encountered option with unsupported type. Only atomic_type…
• Failed to decode schema due to an issue with Field proto
• Unrecognized atomic_type
• Unrecognized atomic_type
• Unrecognized type_info

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/073723cced296d72. Report an issue: GitHub.

Appendix: source

Thrown at sdks/python/apache_beam/typehints/schemas.py:626

      else:
        return row_type.RowTypeConstraint.from_user_type(
            user_type,
            schema_options=schema_options,
            field_options=field_options)

    elif type_info == "logical_type":
      if fieldtype_proto.logical_type.urn == PYTHON_ANY_URN:
        return Any
      else:
        return LogicalType.from_runner_api(
            fieldtype_proto.logical_type).language_type()

    elif type_info == "iterable_type":
      return Sequence[self.typing_from_runner_api(
          fieldtype_proto.iterable_type.element_type)]

    else:
      raise ValueError(f"Unrecognized type_info: {type_info!r}")

  def named_tuple_from_schema(self, schema: schema_pb2.Schema) -> type:
    from apache_beam import coders

    type_name = 'BeamSchema_{}'.format(schema.id.replace('-', '_'))

    subfields = []
    descriptions = {}
    for field in schema.fields:
      try:
        field_py_type = self.typing_from_runner_api(field.type)
        if isinstance(field_py_type, row_type.RowTypeConstraint):
          field_py_type = field_py_type.user_type
      except ValueError as e:
        raise ValueError(
            "Failed to decode schema due to an issue with Field proto:\n\n"
            f"{text_format.MessageToString(field)}") from e

View on GitHub (pinned to 12126d8942)