package aggregate
- Alphabetic
- Public
- All
Type Members
-
class
AggregateHelper extends Serializable
This class is in Java, because constants are used in expressions in editors - see
pl.touk.nussknacker.engine.flink.util.transformer.aggregate.SlidingAggregateTransformerV2
.This class is in Java, because constants are used in expressions in editors - see
pl.touk.nussknacker.engine.flink.util.transformer.aggregate.SlidingAggregateTransformerV2
. and scala objects are not good for that. Be aware that. If you add some new aggregator please add it also there to make sure that it will be available in selectbox.You should define
#AGG
global variable, because it is used in editors. - case class AggregateWindowsConfig(tumblingWindowsOffset: Option[Duration]) extends Product with Serializable
- abstract class Aggregator extends AggregateFunction[AnyRef, AnyRef, AnyRef]
- class AggregatorFunction[MapT[K, V]] extends LatelyEvictableStateFunction[ValueWithContext[StringKeyedValue[AnyRef]], ValueWithContext[AnyRef], MapT[Long, AnyRef]] with AggregatorFunctionMixin[MapT]
- trait AggregatorFunctionMixin[MapT[K, V]] extends RichFunction
- class CardinalityAggregator[Wrapper <: CardinalityWrapper] extends Aggregator with Serializable
-
class
EmitExtraWindowWhenNoDataTumblingAggregatorFunction[MapT[K, V]] extends KeyedProcessFunction[String, ValueWithContext[StringKeyedValue[AnyRef]], ValueWithContext[AnyRef]] with StateHolder[MapT[Long, AnyRef]] with AggregatorFunctionMixin[MapT]
It behaves similar to Flink's TumblingWindow with one difference that we produce extra zero aggregate for each key when no data arrived.
It behaves similar to Flink's TumblingWindow with one difference that we produce extra zero aggregate for each key when no data arrived. For this purpose we've used
AggregatorFunctionMixin
withminimalResolutionMs = timeWindowLengthMillis
. We use it because it handles out of order elements. The other difference from AggregatorFunction is that we emit event only in timer and handle state eviction on ours own. -
class
EmitWhenEventLeftAggregatorFunction[MapT[K, V]] extends LatelyEvictableStateFunction[ValueWithContext[StringKeyedValue[AnyRef]], ValueWithContext[AnyRef], MapT[Long, AnyRef]] with AggregatorFunctionMixin[MapT]
It behaves the same as AggregatorFunction with one difference that also publish events when some event will left the slide.
- class EnrichingWithKeyFunction extends ProcessWindowFunction[AnyRef, ValueWithContext[AnyRef], String, TimeWindow]
- case class HyperLogLogPlusAggregator(p: Int = 10, sp: Int = 15) extends CardinalityAggregator[HyperLogLogPlusWrapper] with Product with Serializable
- class HyperLogLogPlusWrapper extends CardinalityWrapper
-
class
OnEventTriggerWindowOperator[A] extends WindowOperator[String, Input[A], AnyRef, ValueWithContext[AnyRef], TimeWindow]
- Annotations
- @silent( "deprecated" )
- sealed abstract final class SessionWindowTrigger extends Enum[SessionWindowTrigger]
- sealed abstract final class TumblingWindowTrigger extends Enum[TumblingWindowTrigger]
-
class
UnwrappingAggregateFunction[Input] extends AggregateFunction[ValueWithContext[StringKeyedValue[Input]], AnyRef, AnyRef]
This class unwraps value from input's KeyedValue.
This class unwraps value from input's KeyedValue.
NOTE: it would be much cleaner if we evaluated aggregateBy here. However, FLINK-10250 prevents us from doing this and we *have* to compute it beforehand
When using this class it's important that aggregator, passedType and unwrap must match: unwrap result is of passedType and can be processed by aggregator
Value Members
- object AggregateWindowsConfig extends Serializable
- object AggregateWindowsOffsetProvider
- object EnrichingWithKeyFunction extends Serializable
- object OnEventTriggerWindowOperator extends Serializable
- object aggregates
- object sampleTransformers
-
object
transformers
- Annotations
- @PublicEvolving()
- object triggers