apache_beam.transforms.userstate module

User-facing interfaces for the Beam State and Timer APIs.

Experimental; no backwards-compatibility guarantees.

class apache_beam.transforms.userstate.StateSpec

Bases: future.types.newobject.newobject

Specification for a user DoFn state cell.

to_runner_api(context)

class apache_beam.transforms.userstate.BagStateSpec(name, coder)

Bases: apache_beam.transforms.userstate.StateSpec

Specification for a user DoFn bag state cell.

to_runner_api(context)

class apache_beam.transforms.userstate.SetStateSpec(name, coder)

Bases: apache_beam.transforms.userstate.StateSpec

Specification for a user DoFn Set State cell

to_runner_api(context)

class apache_beam.transforms.userstate.CombiningValueStateSpec(name, coder=None, combine_fn=None)

Bases: apache_beam.transforms.userstate.StateSpec

Specification for a user DoFn combining value state cell.

Initialize the specification for CombiningValue state.

CombiningValueStateSpec(name, combine_fn) -> Coder-inferred combining value

state spec.

CombiningValueStateSpec(name, coder, combine_fn) -> Combining value state

spec with coder and combine_fn specified.

Parameters:

• name (str) – The name by which the state is identified.
• coder (Coder) – Coder specifying how to encode the values to be combined. May be inferred.
• combine_fn (CombineFn or callable) – Function specifying how to combine the values passed to state.

to_runner_api(context)

class apache_beam.transforms.userstate.TimerSpec(name, time_domain)

Bases: future.types.newobject.newobject

Specification for a user stateful DoFn timer.

to_runner_api(context)

apache_beam.transforms.userstate.on_timer(timer_spec)

Decorator for timer firing DoFn method.

This decorator allows a user to specify an on_timer processing method in a stateful DoFn. Sample usage:

class MyDoFn(DoFn):
  TIMER_SPEC = TimerSpec('timer', TimeDomain.WATERMARK)

  @on_timer(TIMER_SPEC)
  def my_timer_expiry_callback(self):
    logging.info('Timer expired!')

apache_beam.transforms.userstate.get_dofn_specs(dofn)

Gets the state and timer specs for a DoFn, if any.

Parameters:dofn (apache_beam.transforms.core.DoFn) – The DoFn instance to introspect for timer and state specs.

apache_beam.transforms.userstate.is_stateful_dofn(dofn)

Determines whether a given DoFn is a stateful DoFn.

apache_beam.transforms.userstate.validate_stateful_dofn(dofn)

Validates the proper specification of a stateful DoFn.

class apache_beam.transforms.userstate.RuntimeTimer(timer_spec)

Bases: future.types.newobject.newobject

Timer interface object passed to user code.

clear()

set(timestamp)

class apache_beam.transforms.userstate.RuntimeState

Bases: future.types.newobject.newobject

State interface object passed to user code.

prefetch()

class apache_beam.transforms.userstate.AccumulatingRuntimeState

Bases: apache_beam.transforms.userstate.RuntimeState

read()

add(value)

clear()

class apache_beam.transforms.userstate.BagRuntimeState

Bases: apache_beam.transforms.userstate.AccumulatingRuntimeState

Bag state interface object passed to user code.

class apache_beam.transforms.userstate.SetRuntimeState

Bases: apache_beam.transforms.userstate.AccumulatingRuntimeState

Set state interface object passed to user code.

class apache_beam.transforms.userstate.CombiningValueRuntimeState

Bases: apache_beam.transforms.userstate.AccumulatingRuntimeState

Combining value state interface object passed to user code.

class apache_beam.transforms.userstate.UserStateContext

Bases: future.types.newobject.newobject

Wrapper allowing user state and timers to be accessed by a DoFnInvoker.

get_timer(timer_spec, key, window)

get_state(state_spec, key, window)

commit()