Apache camel 驼峰休息DSL-聚合策略奇怪的行为

标题有点像拉票者，如果它不起作用，当然是我的错，因为它应该起作用

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

我想执行从rdbms到solr和mongo db的数据传输。 为此，我必须完成以下步骤（例如）：

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

• 获取要转移的客户ID
• 获取客户的详细信息
• 获取客户发票
• 向客户付款

然后，聚合并保存到mongo db和solr以进行索引

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

这是我的代码，但我无法让它工作：

from("seda:initial-data-transfer")
        .setProperty("recipientList", simple("direct:details,direct:invoices,direct:payments"))
        .setProperty("afterAggregate", simple("direct:mongodb,direct:solr"))
        .setBody(constant("{{query.initial-data-transfer.ids}}"))
        .to(jdbc)
        .process(new RowSetIdsProcessor())
        .split().tokenize(",", 1000) // ~200k ids - group by 1000 ids
        .to("direct:customers-ids");

from("direct:customers-ids")
        .recipientList(exchangeProperty("recipientList").tokenize(","))
        // ? .aggregationStrategy(new CustomerAggregationStrategy()).parallelProcessing()
        .aggregate(header("CamelCorrelationId"), new CustomerAggregationStrategy())             
        .completionPredicate(new CustomerAggregationPredicate()) // true if details + invoices + payments, etc ....
        // maybe a timeOut here ?
        .process(businessDataServiceProcessor)
        .recipientList(exchangeProperty("afterAggregate").tokenize(","));

from("direct:details")
        .setHeader("query", constant("{{query.details}}"))
        .bean(SqlTransform.class,"detailsQuery").to(jdbc)
        .process(new DetailsProcessor());

from("direct:invoices")
        .setHeader("query", constant("{{query.invoices}}"))
        .bean(SqlTransform.class,"invoicessQuery").to(jdbc)
        .process(new InvoicesProcessor());

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

我不明白聚合策略是如何工作的。 有时，我可以执行2或3个1000个ID的块，并保存到mongo DB和Solr，但之后，aggregationStrategy中的所有交换都是空的

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

我试了很多想法。。但每次聚合都会失败

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

谢谢你的帮助

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

更新：

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

以下是CustomerAggregationStrategy的一部分：

public class CustomerAggregationStrategy implements AggregationStrategy {
    @Override
    public Exchange aggregate(Exchange oldExchange, Exchange newExchange) {
        Message newIn = newExchange.getIn();

        CustomerDataCollector collector = null;
        if (oldExchange == null) {
            int completionSize = newExchange.getProperty("completionSize", Integer.class);
            collector = new CustomerDataCollector(completionSize);
            CollectData(collector, newIn, newExchange);
            newIn.setBody(collector);
            return newExchange;
        }
        collector = oldExchange.getIn().getBody(CustomerDataCollector.class);
        CollectData(collector, newIn, newExchange);
        return oldExchange;
    }

    private void CollectData(CustomerDataCollector collector, Message message, Exchange exchange) {
        String recipientListEndpoint = (String)exchange.getProperty(Exchange.RECIPIENT_LIST_ENDPOINT);
        switch (recipientListEndpoint){
            case "direct://details" :
                collector.setDetails(message.getBody(Map.class));
                break;
            case "direct://invoices" :
                collector.setInvoices(message.getBody(Map.class));
                break;
            case "direct://payments" :
                collector.setPayments(message.getBody(Map.class));
                break;
        }
    }
}

String camelCorrelationId = (String)exchange.getProperty(Exchange.CORRELATION_ID);

[t-AggregateTask] .i.c.a.CustomerAggregationStrategy : CustomerAggregationStrategy.CollectData : direct://details  ID-UC-0172-50578-1484523575668-0-5
[t-AggregateTask] .i.c.a.CustomerAggregationStrategy : CustomerAggregationStrategy.CollectData : direct://invoices ID-UC-0172-50578-1484523575668-0-5
[t-AggregateTask] .i.c.a.CustomerAggregationStrategy : CustomerAggregationStrategy.CollectData : direct://payments ID-UC-0172-50578-1484523575668-0-5

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

更新：

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

我可以将此记录在CustomerAggregationStrategy中：

public class CustomerAggregationStrategy implements AggregationStrategy {
    @Override
    public Exchange aggregate(Exchange oldExchange, Exchange newExchange) {
        Message newIn = newExchange.getIn();

        CustomerDataCollector collector = null;
        if (oldExchange == null) {
            int completionSize = newExchange.getProperty("completionSize", Integer.class);
            collector = new CustomerDataCollector(completionSize);
            CollectData(collector, newIn, newExchange);
            newIn.setBody(collector);
            return newExchange;
        }
        collector = oldExchange.getIn().getBody(CustomerDataCollector.class);
        CollectData(collector, newIn, newExchange);
        return oldExchange;
    }

    private void CollectData(CustomerDataCollector collector, Message message, Exchange exchange) {
        String recipientListEndpoint = (String)exchange.getProperty(Exchange.RECIPIENT_LIST_ENDPOINT);
        switch (recipientListEndpoint){
            case "direct://details" :
                collector.setDetails(message.getBody(Map.class));
                break;
            case "direct://invoices" :
                collector.setInvoices(message.getBody(Map.class));
                break;
            case "direct://payments" :
                collector.setPayments(message.getBody(Map.class));
                break;
        }
    }
}

String camelCorrelationId = (String)exchange.getProperty(Exchange.CORRELATION_ID);

[t-AggregateTask] .i.c.a.CustomerAggregationStrategy : CustomerAggregationStrategy.CollectData : direct://details  ID-UC-0172-50578-1484523575668-0-5
[t-AggregateTask] .i.c.a.CustomerAggregationStrategy : CustomerAggregationStrategy.CollectData : direct://invoices ID-UC-0172-50578-1484523575668-0-5
[t-AggregateTask] .i.c.a.CustomerAggregationStrategy : CustomerAggregationStrategy.CollectData : direct://payments ID-UC-0172-50578-1484523575668-0-5

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

CamelCorrelationId的值与预期值相同。 我想CamelCorrelationId是正确的。是吗？

好的，现在好多了

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

在tokeniszer之后，我像这样设置属性CustomCorrelationId

    .split().tokenize(",", 1000)
    .setProperty("CustomCorrelationId",header("breadcrumbId"))
    .to("direct:customers-ids")

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

并在此值上进行如下聚合：

    from("direct:customers-ids")
            .recipientList(exchangeProperty("recipientList").tokenize(","))

    from("direct:details")
            .setHeader("query", constant("{{query.details}}"))
            .bean(SqlTransform.class,"detailsQuery").to(jdbc)
            .process(new DetailsProcessor())
            .to("direct:aggregate");      

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));

这项工作现在很好，数据已正确聚合。谢谢你的帮助。

---

您指出了方向。

如何设置标题

CamelCorrelationId

？相关性表达式必须是您在交换之间的共同点，我认为CorrelationId总是不同的，那么它就不会像这样工作。。。。如果您可以为CustomerAggregationStrategy类添加代码（或至少提及逻辑），这将有助于更好地解决路由问题。我不会根据以下内容设置CamelCorrelationID：“某些EIP模式将派生一个子消息，在这些情况下，Camel将在Exchange上添加一个相关id，作为其键Exchange.correlation_id的属性，该键链接回源Exchange。例如，拆分器、多播、收件人列表和有线点击EIP就是这样做的。”。RecipientList EIP就是这样做的。我用一些代码更新了这个问题。谢谢你的帮助。是的，那么它将永远不会聚合，因为所有邮件的correlationId都不同。首先，你必须定义“公共密钥”“您将在其上聚合，然后，该策略定义如何合并两个聚合的消息。例如，如果您希望聚合每个客户的所有消息，那么correlationExpression必须类似于

头（customerId）

@ruff查看日志。CamelCorrelationId是相同的。这是巧合吗？

    from("direct:aggregate").routeId("aggregate")
            .log("route : ${routeId}")
            .aggregate(property("CustomCorrelationId"), new CustomAggregationStrategy())
            .completionPredicate(new CustomerAggregationPredicate())
            .process(businessDataServiceProcessor)
            .recipientList(exchangeProperty("afterAggregate").tokenize(","));