Java多线程只在第一次完成时调用Other

Question

ExecutorService executor = Executors.newFixedThreadPool(10);
Runnable firstWorker = new DataComparison(DataComparison.FIRST_THREAD_INSTANCE,args[0], args[1], args[2], runSkey, fic_mis_date,legal_entity, args[6],failover);
executor.execute(firstWorker);
Runnable secondWorker = new DataComparison(DataComparison.SECOND_THREAD_INSTANCE,args[0], args[1], args[2], runSkey, fic_mis_date,legal_entity, args[6],failover);
executor.execute(secondWorker);
Runnable thirdWorker = new DataComparison(DataComparison.THIRD_THREAD_INSTANCE,args[0], args[1], args[2], runSkey, fic_mis_date,legal_entity, args[6],failover);
executor.execute(thirdWorker);

这是我的代码，其中我想只在第一个线程完成处理时调用第二个和第三个线程。我分不清AwaitTermination和sleep。请建议。

Answer 1

使用Java 8 CompletableFuture你可以写

 ExecutorService executor = Executors.newFixedThreadPool(10);
 Runnable firstWorker = new DataComparison(DataComparison.FIRST_THREAD_INSTANCE,args[0], args[1], args[2], runSkey, fic_mis_date,legal_entity, args[6],failover);

 Runnable secondWorker = new DataComparison(DataComparison.SECOND_THREAD_INSTANCE,args[0], args[1], args[2], runSkey, fic_mis_date,legal_entity, args[6],failover);

  Runnable thirdWorker = new DataComparison(DataComparison.THIRD_THREAD_INSTANCE,args[0], args[1], args[2], runSkey, fic_mis_date,legal_entity, args[6],failover);
 CompletableFuture firstTask = CompletableFuture.runAsync(firstWorker,
          executor);
 firstTask.thenRunAsync(secondWorker,executor)
 firstTask.thenRunAsync(thirdWorker,executor);

这将在第一个任务完成时启动第二个和第三个任务。

Answer 2

如果我对你的问题理解正确的话，使用Guava对期货的支持可能真的很有帮助。我会参考谷歌的这篇文章来学习更多关于ListenableFuture的知识。下面是基于您的代码的代码示例，但是如果您确实计划使用Guava的Futures，我可能会以稍微不同的方式构建代码:

    ListeningExecutorService s = MoreExecutors.listeningDecorator(Executors.newFixedThreadPool(10));
    Runnable firstWorker = new DataComparison(1);
    final Runnable secondWorker = new DataComparison(2);
    final Runnable thirdWorker = new DataComparison(3);
    ListenableFuture f = s.submit(firstWorker);
    Futures.addCallback(f, new FutureCallback() {
        @Override
        public void onSuccess(Object o) {
            secondWorker.run();
        }
        @Override
        public void onFailure(Throwable throwable) {
        }
    });
    Futures.addCallback(f, new FutureCallback() {
        @Override
        public void onSuccess(Object o) {
            thirdWorker.run();
        }
        @Override
        public void onFailure(Throwable throwable) {
        }
    });

这里发生的事情是，我们首先提交第一个worker，然后向它添加2个回调(secondWorker和thirdWorker)。一旦firstWorker成功完成，就会调用onSuccess方法。如果抛出异常，则调用onFailure。

Answer 3

这是一个简单的可能性。你还可以做其他更复杂的事情，但如果我正确理解了你的问题，这应该是你想要的:

final Runnable firstWorker = new DataComparison(DataComparison.FIRST_THREAD_INSTANCE,args[0], args[1], args[2], runSkey, fic_mis_date,legal_entity, args[6],failover);
final Runnable secondWorker = new DataComparison(DataComparison.SECOND_THREAD_INSTANCE,args[0], args[1], args[2], runSkey, fic_mis_date,legal_entity, args[6],failover);
final Runnable thirdWorker = new DataComparison(DataComparison.THIRD_THREAD_INSTANCE,args[0], args[1], args[2], runSkey, fic_mis_date,legal_entity, args[6],failover);
Runnable combined = new Runnable() {
  @Override
  public void run() {
    firstWorker.run();
    executor.execute(secondWorker);
    executor.execute(thirdWorker);
  }
};
executor.execute(combined);