Java 如何从Akka Streams接收器中抛出的异常中恢复？

如何从Akka Streams接收器中抛出的异常中恢复

简单的例子：

    Source<Integer, NotUsed> integerSource = Source.from(Arrays.asList(1, 2, 3, 4, 5, 6, 7, 8, 9));

    integerSource.runWith(Sink.foreach(x -> {
      if (x == 4) {
        throw new Exception("Error Occurred");
      }
      System.out.println("Sink: " + x);
    }), system);

我如何处理异常并从源代码转到下一个元素？aka 5,6,7,8,9默认情况下，在抛出异常时停止流。要更改监视策略以删除导致异常的消息并继续处理下一条消息，请使用恢复策略。例如：

final Function<Throwable, Supervision.Directive> decider =
  exc -> {
    return Supervision.resume();
  };

final Sink<Integer, CompletionStage<Done>> printSink =
  Sink.foreach(x -> {
    if (x == 4) {
      throw new Exception("Error Occurred");
    }
    System.out.println("Sink: " + x);
  });

final RunnableGraph<CompletionStage<Done>> runnableGraph =
  integerSource.toMat(printSink, Keep.right());

final RunnableGraph<CompletionStage<Done>> withResumingSupervision =
  runnableGraph.withAttributes(ActorAttributes.withSupervisionStrategy(decider));

final CompletionStage<Done> result = withResumingSupervision.run(system);

您还可以为不同类型的异常定义不同的监督策略：

final Function<Throwable, Supervision.Directive> decider =
  exc -> {
    if (exc instanceof MySpecificException) return Supervision.resume();
    else return Supervision.stop();
  };

默认情况下，当抛出异常时，停止流。要更改监视策略以删除导致异常的消息并继续处理下一条消息，请使用恢复策略。例如：

final Function<Throwable, Supervision.Directive> decider =
  exc -> {
    return Supervision.resume();
  };

final Sink<Integer, CompletionStage<Done>> printSink =
  Sink.foreach(x -> {
    if (x == 4) {
      throw new Exception("Error Occurred");
    }
    System.out.println("Sink: " + x);
  });

final RunnableGraph<CompletionStage<Done>> runnableGraph =
  integerSource.toMat(printSink, Keep.right());

final RunnableGraph<CompletionStage<Done>> withResumingSupervision =
  runnableGraph.withAttributes(ActorAttributes.withSupervisionStrategy(decider));

final CompletionStage<Done> result = withResumingSupervision.run(system);

您还可以为不同类型的异常定义不同的监督策略：

final Function<Throwable, Supervision.Directive> decider =
  exc -> {
    if (exc instanceof MySpecificException) return Supervision.resume();
    else return Supervision.stop();
  };

---

是否要跳过特定项目？然后过滤它们，查看更多细节：我不知道哪个元素会导致异常。我的接收器可能是Kafka或RabbitMq，因此如果说到rabbbitmq的连接暂时失败，那么我不想停止流，而是为未来的元素继续。是否要跳过特定的项？然后过滤它们，查看更多细节：我不知道哪个元素会导致异常。我的接收器可能是Kafka或RabbitMq，因此，如果说到rabbbitmq的连接暂时失败，那么我不想停止流，而是为将来的元素继续谢谢Jeff。有一个小小的修正：withResumingSupervision应该是Runnablegraphy类型谢谢Jeff。它做了一个小小的修正：恢复监督应为RunnableGraph类型