apache/beam · error · ValueError

No existing_windows in this context.

Error message

No existing_windows in this context.

What it means

The sibling of the timestamp check: ProcessContext.existing_windows unconditionally raises ValueError('No existing_windows in this context.') because this context type never carries window information. Any code path that asks for windows through this context is invalid by construction.

Source

Thrown at sdks/python/apache_beam/runners/common.py:1905

class _NoContext(WindowFn.AssignContext):
  """An uninspectable WindowFn.AssignContext."""
  NO_VALUE = object()

  def __init__(self, value, timestamp=NO_VALUE):
    self.value = value
    self._timestamp = timestamp

  @property
  def timestamp(self):
    if self._timestamp is self.NO_VALUE:
      raise ValueError('No timestamp in this context.')
    else:
      return self._timestamp

  @property
  def existing_windows(self):
    raise ValueError('No existing_windows in this context.')


class DoFnState(object):
  """For internal use only; no backwards-compatibility guarantees.

  Keeps track of state that DoFns want, currently, user counters.
  """
  def __init__(self, counter_factory):
    self.step_name = ''
    self._counter_factory = counter_factory

  def counter_for(self, aggregator):
    """Looks up the counter for this aggregator, creating one if necessary."""
    return self._counter_factory.get_aggregator_counter(
        self.step_name, aggregator)


# TODO(robertwb): Replace core.DoFnContext with this.

View on GitHub (pinned to 12126d8942)

Solutions

1. Access windows through the correct mechanism: declare the DoFn with the windows-related context (e.g. use the `window` DoFn parameter via beam.DoFn.ProcessContext or @beam.DoFn.window_param-style access / process(ctx) with windowed value).
2. In tests, wrap inputs in WindowedValue so window info is available through the runner path.
3. Refactor to not need input windows; use beam.WindowInto or window-aware transforms instead.

Example fix

// before
class MyDoFn(beam.DoFn):
    def process(self, ctx):
        print(ctx.existing_windows)  # always raises
// after
class MyDoFn(beam.DoFn):
    def process(self, element, window=beam.DoFn.WindowParam):
        print(window)  # proper window access

if not isinstance(ctx, type(None)) and getattr(ctx, 'existing_windows', None) is None:
    skip_window_logic()

def has_windows(ctx):
    try:
        _ = ctx.existing_windows
        return True
    except ValueError:
        return False

try:
    windows = ctx.existing_windows
except ValueError:
    windows = [GlobalWindow()]  # conservative default
    log.warning('Window info unavailable; assuming global window')

When it happens

Trigger: Accessing `ctx.existing_windows` on a ProcessContext (e.g. inside a DoFn's process via the process context or in tests) when the runner-provided context does not expose windows.

Common situations: Trying to read input windows in a DoFn without the proper context (e.g. via DoFnTester or a plain Context rather than the windowed receiver); assuming windows are always available in ctx.

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

• No timestamp in this context.
• Pipeline has validations errors:
• Found more than one renderer for option: %s
• The given pcoll %s is not a dict, an iterable or a PCollecti
• All PCollections must belong to the same pipeline.

AI-assisted analysis of apache/beam@12126d8942 (2026-09-13). Data as JSON: /api/errors/60ebb77abc9b2d03. Report an issue: GitHub.