apache/beam · error · RuntimeError

Columns are not specified. Please specify the column for…

Error message

Columns are not specified. Please specify the column for the  op %s

What it means

All TFT transform op base classes (ApplyTransforms base in tft.py) require a non-empty columns list identifying which PCollection columns the transform applies to. The __init__ raises a RuntimeError when columns is falsy (None, empty list), because apply_transform would otherwise have no target column.

Solutions

Pass the target column name(s): e.g. ScaleToZScore(columns=['feature_1']).
Verify the source of the columns list is non-empty before constructing configs.
Validate column names against the PCollection schema so downstream key errors are also avoided.

Example fix

# before
transform = tft.ScaleToZScore()  # columns missing

# after
transform = tft.ScaleToZScore(columns=['age'])

Defensive patterns

Strategy: validation

Validate before calling

def make_scale_zscore(columns):
    if not columns:
        raise ValueError('columns must be a non-empty list of column names')
    return tft.ScaleToZScore(columns=columns)

Type guard

def has_columns(columns) -> bool:
    return isinstance(columns, (list, tuple)) and len(columns) > 0 and all(isinstance(c, str) for c in columns)

Try / catch

try:
    transforms = [tft.ScaleToZScore(columns=cols)]
except RuntimeError as e:
    if 'Columns are not specified' in str(e):
        raise ValueError(f'Provide target columns for transform: {e}') from e
    raise

Prevention

Always pass columns as the first argument to every TFT transform config.
Validate the transform-config list (non-empty columns) before building MLTransform.
Use keyword arguments (columns=[...]) to avoid positional mix-ups.
Add a schema test asserting each config targets an existing column.

When it happens

Trigger: Constructing a TFT transform config like ScaleToZScore(), ScaleMinMax(), ComputeAndApplyVocab() with columns=None or columns=[] (or omitting the first positional argument).

Common situations: Building transform configs programmatically from empty config lists, copy-pasting a transform instantiation without filling in column names, or a config loader returning an empty columns key.

Understand the failure class

Background: "missing required argument" and "the following required arguments were not provided": what required-argument errors mean and how to fix them — this error's family across 20 libraries.

Related errors

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/61abaa4760db5e6b. Report an issue: GitHub.

Appendix: source

Thrown at sdks/python/apache_beam/ml/transforms/tft.py:97


# TODO: https://github.com/apache/beam/pull/29016
# Add support for outputting artifacts to a text file in human readable form.
class TFTOperation(BaseOperation[common_types.TensorType,
                                 common_types.TensorType]):
  def __init__(self, columns: list[str]) -> None:
    """
    Base Operation class for TFT data processing transformations.
    Processing logic for the transformation is defined in the
    apply_transform() method. If you have a custom transformation that is not
    supported by the existing transforms, you can extend this class
    and implement the apply_transform() method.
    Args:
      columns: List of column names to apply the transformation.
    """
    super().__init__(columns)
    if not columns:
      raise RuntimeError(
          "Columns are not specified. Please specify the column for the "
          " op %s" % self.__class__.__name__)

  def get_ptransform_for_processing(self, **kwargs) -> beam.PTransform:
    from apache_beam.ml.transforms.handlers import TFTProcessHandler
    params = {}
    artifact_location = kwargs.get('artifact_location')
    if not artifact_location:
      raise RuntimeError(
          "artifact_location is not specified. Please specify the "
          "artifact_location for the op %s" % self.__class__.__name__)

    artifact_mode = kwargs.get('artifact_mode')
    if artifact_mode:
      params['artifact_mode'] = artifact_mode
    return TFTProcessHandler(artifact_location=artifact_location, **params)

  @tf.function

View on GitHub (pinned to 12126d8942)