apache/beam · error · ValueError

Unknown spec type ' ' in

Error message

Unknown spec type '{spec.type}' in {spec}

What it means

In Specifiable.from_spec, after the Spec has a type and the type's subspace is found, the concrete subclass is looked up in _KNOWN_SPECIFIABLE[subspace]. If the type is in a subspace key space but not actually registered there (or lookup misses), a ValueError 'Unknown spec type ...' is raised. It indicates the type string is not a registered Specifiable class.

Solutions

1. Correct Spec.type to the exact registered class name string.
2. Register the custom class with the @specifiable decorator before from_spec is called.
3. Confirm the module containing the class is imported so registration executes.
4. Regenerate/re-serialize old pipeline specs against the current Beam version's registry.

Example fix

// before
@dataclass
class MyDetector: ...  # never registered
Spec(type="MyDetector")
// after
@specifiable("MyDetector", Subspace.DETECTOR)
class MyDetector(BaseDetector): ...
Spec(type="MyDetector")

known = set().union(*_KNOWN_SPECIFIABLE.values())
if spec.type not in known:
    raise ValueError(f"unknown spec type: {spec.type!r}; known: {sorted(known)}")

def is_registered_spec(spec): return spec.type in set().union(*_KNOWN_SPECIFIABLE.values())

try:
    obj = Specifiable.from_spec(spec)
except ValueError as e:
    logger.error("unknown spec type in %s: %s", spec, e)
    raise

When it happens

Trigger: Passing a Spec whose .type string is not present in the registry for its subspace - misspelled built-in names, unregistered custom classes, or types from a different Beam version.

Common situations: Typos like 'z_score' vs 'ZScore'; custom detectors not decorated with @specifiable; pipeline specs serialized under an older Beam release whose registry differed; stale pickled Spec objects after upgrade.

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

• subspace for not found.
• Spec type not found in
• Coder registry has no fallback coder. This can happen if…
• "IQR.learn_one expected univariate input, but got
• "IQR.score_one expected univariate input, but got

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/b75ab52fad7d9737. Report an issue: GitHub.

Appendix: source

Thrown at sdks/python/apache_beam/ml/anomaly/specifiable.py:175

                spec: Spec,
                _run_init: bool = True) -> Union[Self, type[Self]]:
    """Generate a `Specifiable` subclass object based on a spec.

    Args:
      spec: the specification of a `Specifiable` subclass object
      _run_init: whether to call `__init__` or not for the initial instantiation

    Returns:
      Self: the `Specifiable` subclass object
    """
    if spec.type is None:
      raise ValueError(f"Spec type not found in {spec}")

    subspace = _spec_type_to_subspace(spec.type)
    subclass: type[Self] = _KNOWN_SPECIFIABLE[subspace].get(spec.type, None)

    if subclass is None:
      raise ValueError(f"Unknown spec type '{spec.type}' in {spec}")

    if spec.config is None:
      # when functions or classes are used as arguments, we won't try to
      # create an instance.
      return subclass

    kwargs = {
        k: _specifiable_from_spec_helper(v, _run_init)
        for k, v in spec.config.items()
    }

    if _run_init:
      kwargs["_run_init"] = True
    return subclass(**kwargs)

  def to_spec(self) -> Spec:
    """Generate a spec from a `Specifiable` subclass object.

View on GitHub (pinned to 12126d8942)