apache/beam · error · WontImplementError

Unsupported value for new column

Error message

Unsupported value for new column '{name}': '{value}'. Only callables and DeferredSeries instances are supported. Other types make this operation sensitive to the order of the data

What it means

DeferredDataFrame.assign rejects column values that are neither callables nor DeferredSeries. Scalars, lists, or plain pandas Series passed as a new column value would need to be broadcast positionally, which depends on the order of rows in the distributed data, so Beam raises WontImplementError with reason 'order-sensitive'.

Solutions

Pass a callable: ddf.assign(col=lambda df: <DeferredSeries expression>).
Wrap array-like values as DeferredSeries first, e.g. via beam.dataframe or DeferredSeries.from_monotonic... (construct from the same pipeline).
Compute constant columns with an elementwise expression over an existing column, e.g. lambda df: df.existing * 0 + value.

Example fix

// before
ddf.assign(total=[1, 2, 3])

// after
ddf.assign(total=lambda df: df['other_col'] * 0 + 1)  # or any callable returning DeferredSeries

Defensive patterns

Strategy: validation

Validate before calling

for name, value in assign_kwargs.items():
    if not callable(value) and not isinstance(value, DeferredSeries):
        assign_kwargs[name] = lambda df, v=value: df[ref_col] * 0 + v  # or wrap as DeferredSeries

Type guard

def is_assignable(value) -> bool:
    return callable(value) or isinstance(value, DeferredSeries)

Try / catch

from apache_beam.dataframe import frame_base
try:
    out = ddf.assign(**kwargs)
except frame_base.WontImplementError as e:
    # convert offending columns to callables and retry
    out = ddf.assign(**{k: (v if callable(v) or isinstance(v, DeferredSeries) else (lambda df, c=v: df[df.columns[0]] * 0 + c)) for k, v in kwargs.items()})

Prevention

Always pass callables to assign in Beam DataFrame code.
Never pass positional lists/Series as assign values in distributed contexts.
Review ported notebook code for scalar/list assign patterns.

When it happens

Trigger: Calling `ddf.assign(col=<scalar | list | pd.Series | np.ndarray>)` — any kwargs value that is not callable and not a DeferredSeries.

Common situations: Porting `df.assign(total=[1,2,3])` or `df.assign(flag=0)` from a notebook to a Beam DataFrame transform; assigning a plain pandas Series column to a distributed frame.

Understand the failure class

Background: UnsupportedOperationException and "is not supported" errors: when a library deliberately refuses a call — this error's family across 30 libraries.

Related errors

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/dd3952e8a77df3fe. Report an issue: GitHub.

Appendix: source

Thrown at sdks/python/apache_beam/dataframe/frames.py:2731


  @property  # type: ignore
  @frame_base.with_docs_from(pd.DataFrame)
  def axes(self):
    return (self.index, self.columns)

  @property  # type: ignore
  @frame_base.with_docs_from(pd.DataFrame)
  def dtypes(self):
    return self._expr.proxy().dtypes

  @frame_base.with_docs_from(pd.DataFrame)
  def assign(self, **kwargs):
    """``value`` must be a ``callable`` or :class:`DeferredSeries`. Other types
    make this operation order-sensitive."""
    for name, value in kwargs.items():
      if not callable(value) and not isinstance(value, DeferredSeries):
        raise frame_base.WontImplementError(
            f"Unsupported value for new column '{name}': '{value}'. Only "
            "callables and DeferredSeries instances are supported. Other types "
            "make this operation sensitive to the order of the data",
            reason="order-sensitive")
    return self._elementwise(
        lambda df, *args, **kwargs: df.assign(*args, **kwargs),
        'assign',
        other_kwargs=kwargs)

  @frame_base.with_docs_from(pd.DataFrame)
  @frame_base.args_to_kwargs(pd.DataFrame)
  @frame_base.populate_defaults(pd.DataFrame)
  def explode(self, column, ignore_index):
    # ignoring the index will not preserve it
    preserves = (partitionings.Singleton() if ignore_index
                 else partitionings.Index())
    return frame_base.DeferredFrame.wrap(
        expressions.ComputedExpression(

View on GitHub (pinned to 12126d8942)