Just Spring Integration

Aggregators

The job of Aggregators is to assemble multiple messages to create a single parent message. They are the opposite of splitters, since they require information to begin and end a task. They have to maintain state for this reason. They also need to follow certain strategies to correlate the messages and release them after they are aggregated.

This is a complex task because all the messages of a set have to arrive before the aggregators can start work. Before we look at a complicated aggregation task, let’s consider a simple task based on default correlation and release strategies.

The TradeAggregator defined below is a simple aggregator whose job is to aggregate the incoming child Trades into a parent Trade.

public class TradeAggregator {

  public ITrade aggregateTrade(List<ITrade> childTrades) {
    ...
  }
}

This aggregator is declared using the following metadata:

<int:aggregator input-channel="in-channel" 
                output-channel="agg-channel"
                ref="tradeAggregator" 
                method="aggregateTrade">
</int:aggregator> 

<bean id="tradeAggregator"
  class="com.madhusudhan.jsi.flow.aggregator.TradeAggregator" />

<!-- Splitter that would cut the messages for aggregator to re-build -->
<int:splitter input-channel="in-channel" ref="customSplitter"
  output-channel="out-channel">
  </int:splitter>
 <bean id="customSplitter"
  class="com.madhusudhan.jsi.flow.splitter.CustomEncryptedTradeSplitter" />

The ref and method attributes define a specific method on the POJO to be invoked whenever the release strategy is fulfilled. ...