java多线程isalive与否

时间:2012-09-28 10:14:43

标签: java multithreading

这样我生成线程,但在for循环之后如何检查所有线程是否存活 进一步的过程取决于所有的线程
如果线程为dead,则进行进一步处理

for (int i = 0;i<List.size();i++) {
                if(List.get(i).trim().length() >0 ){

                    Thread thread = new Thread("Thread"+i){
                        public void run(){              
                            threading(List.get(i),ctx,userSesionObj);
                        }
                      };
                      thread.start();
                }
            }

这里threading是我的函数,它执行我传递的不同数据

7 个答案:

答案 0 :(得分:4)

您可以使用ExecutorService。事实上,我强烈推荐这一点。您可以指定不同的线程行为等。执行程序将为每个提交的线程返回一个Future对象,然后您可以遍历每个Future收集结果。以下是:

for (Future f : futures) {
   f.get(); // get a result here...
}
当所有Futures都返回数据时(即线程完成时),

将完成。 Executor API的工作级别高于join() / notify()等,使用起来更安全。

有很多功能可以在这里使用,我会推荐tutorials

答案 1 :(得分:2)

您可能希望join线程。参看http://docs.oracle.com/javase/tutorial/essential/concurrency/join.html。鉴于您可以暂停主线程,即。

答案 2 :(得分:1)

使用ExecutorService

- Executors管理Thread个对象和Async任务井

- 这里不是直接运行线程的客户端,而是中间对象。

- 将此与submit()方法结合使用,该方法返回Future对象。

- Future是异步任务的结果

- 您可以使用Future的get()方法,其具有阻止性质,或get()方法以及isDone()方法方法。

答案 3 :(得分:0)

使用Thread类中的方法 isAlive

来自文档:

  

测试此线程是否存活。如果一个线程已经启动并且还没有死亡,那么它就是活着的。

答案 4 :(得分:0)

正如彼得所说,这应该做......

int count = selectedCategoryList.size();
List<Thread> threads = new ArrayList<Thread>();
for (int i = 0; i < count; i++) {
  if(selectedCategoryList.get(i).trim().length() >0 ){
    final Long CategoryID = Long.parseLong(selectedCategoryList.get(i));
    Thread thread = new Thread("ReplaceMedicationThread"+i){
                          public void run(){              
                            threadingForReplacingMedication(CategoryID,ctx,userSesionObj);
                          }
                    };
    threads.add(thread):
   }
}

//start them in a loop
//join them in a loop
//move on....

答案 5 :(得分:0)

在继续处理之前等待一些线程完成的简单方法是通过CompletionService。这就是你如何使用它:

// list of data that need to be processed before continuing execution
List<DataForProcessing> workList = // ...

// An instance of an Executor http://docs.oracle.com/javase/1.5.0/docs/api/java/util/concurrent/Executors.html
Executor exec = // ...

// asume that ResultData is a data structure returned from the processing each
// DataForProcessing instance. If processing does not return a value you could
// parameterise the completion service as CompletionService<Void>.
CompletionService<ResultData> service = new ExecutorCompletionService<ResultData>(executor);

// Submit the tasks for processing. Each task will be processed by the executor and
// its result will be wrapped in a Future and place in a BlockingQueue maintained by
// the CompletionService instance. Again, if your submitted task does not return
// a value you should parameterise your task as Callable<Void>.
for (DataForProcessing data: workList) {
  service.submit(new Callable<ResultData>() {
    public ResultData call() {
      // process the data here and return some result (optionally).
    }
  });
 }

// retrieve the results as the become available.
for (int i = 0; i < workLoad.size(); i++) {
  // Retrieve the next processed result. CompletionService returns each task in a
  // Future which provides you with the means to control the lifecycle of the task.
  // When a submitted task is processed by the executor, its Future is placed on a
  // BlockingQueue maintained by the CompletionService. Invoking take() will return
  // the next available result. If your tasks do not return a result you would still
  // need to call Future<Void> f = service.take() in order to wait for every task
  // to finish before proceeding.
  Future<ResultData> f = service.take();
  // If your task returns some result you may access it via the get() method on the
  // task's Future.
  ResultData result = f.get();
  // do some processing if needed
  process(result);
}

答案 6 :(得分:0)

我刚刚使用简单的一个全局int参数完成了它。 我喜欢这样:

do{
    Thread.sleep(800);
    System.out.println("threadCounter=="+threadCounter);
    }while(threadCounter != 0);

其中threadCounter定义全局int初始化zero

关于制作每个帖子我做threadCounter;

的+1

在我的函数threading中,我在threadCounter

的-1

这里Thread.sleep(800)因为没有运行do-while循环继续。