Package org.voltdb.stream.function

Class BatchAccumulator<I,O>

java.lang.Object

org.voltdb.stream.function.BatchAccumulator<I,O>

Type Parameters:

I - the type of input elements to be accumulated

O - the type of the accumulated result

All Implemented Interfaces:

VoltFunction<I,O>, VoltLifecycle

public class BatchAccumulator<I,O>
extends Object
implements VoltFunction<I,O>

A processing function that accumulates incoming data items into a single aggregated result and emits the result once a batch is processed.

It uses a provided BiFunction to accumulate each input item into an accumulator of type O. After processing a batch of inputs, the accumulated result is emitted through the provided Consumer.

Usage Example

 // Creating an BatchAccumulator instance
 var batchAccumulator = new BatchAccumulator<Integer, Long>(0L, Long::sum);

 // Use it as a function in a pipeline.
 stream
   .withName("simple stream")
   .consumeFromSource(Sources.collection("A", "AB", "CDE"))
   .processWith(input -> input.getBytes(StandardCharsets.UTF_8).length)
   .processWith(batchAccumulator)
   .terminateWithSink(Sinks.consume(output -> log.info("bytes consumed by the batch is {}", output)));
 

Constructor Summary

Constructors

Constructor	Description
BatchAccumulator(O accumulator, BiFunction<O,I,O> accumulatorFunction)	Constructs a BatchAccumulator with the specified initial accumulator and accumulation function.

Method Summary

Modifier and Type	Method	Description
void	batchProcessed(long batchId)	the callback is invoked when the function finishes processing a batch of data
void	initialize(Consumer<O> consumer)	the callback is invoked when the function is created and a consumer is assigned but the function is not yet scheduled to handle any incoming events.
void	process(I input, Consumer<O> consumer, ExecutionContext context)	processes input and emits output messages to a consumer

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface org.voltdb.stream.api.pipeline.VoltFunction

nextBatchStarts

Methods inherited from interface org.voltdb.stream.api.pipeline.VoltLifecycle

configure, destroy

Constructor Details

BatchAccumulator

public BatchAccumulator(O accumulator,
 BiFunction<O,I,O> accumulatorFunction)

Constructs a BatchAccumulator with the specified initial accumulator and accumulation function.

The accumulator function defines how each input item is combined with the current accumulator to produce a new accumulator value.

Parameters:

accumulator - the initial value of the accumulator

accumulatorFunction - a BiFunction that combines the current accumulator with an input item

Method Details

initialize

public void initialize(Consumer<O> consumer)

Description copied from interface: VoltFunction

the callback is invoked when the function is created and a consumer is assigned but the function is not yet scheduled to handle any incoming events.

Specified by:

initialize in interface VoltFunction<I,O>

Parameters:

consumer - a downstream consumer used to emit transformed events to it

process

public void process(I input,
 Consumer<O> consumer,
 ExecutionContext context)

Description copied from interface: VoltFunction

processes input and emits output messages to a consumer

Specified by:

process in interface VoltFunction<I,O>

Parameters:

input - message

consumer - binds this function with next step

context - of the execution

batchProcessed

public void batchProcessed(long batchId)

Description copied from interface: VoltFunction

the callback is invoked when the function finishes processing a batch of data

Specified by:

batchProcessed in interface VoltFunction<I,O>

Parameters:

batchId - globally unique id to track messages related to current batch of data