completionService的使用

来源:互联网 发布:淘宝导航栏代码生成器 编辑:程序博客网 时间:2024/06/05 22:42

CompletionService

接口CompletionService 的功能是以异步的方式一边生产新的任务,一边处理已完成任务的结果,这样可以将执行任务与从处理任务分离开来进行处理。使用submit执行任务,使用take取得已完成的任务,并按照完成这些任务的时间处理它们的结果。

接口CompletionService仅有一个实现类 ExecutorCompletionService,构造方法

public ExecutorCompletionService(Executor executor)public ExecutorCompletionService(Executor executor,                                     BlockingQueue<Future<V>> completionQueue)

需要依赖于Executor对象,大部分的实现也就是使用线程池ThreadPoolExecutor对象。

Future具有阻塞同步性,接口CompletionService可以解决解决这个问题

public class MyCallable implements Callable<String> {    private String userName;    private long sleepValue;    public MyCallable(String userName, long sleepValue) {        this.userName = userName;        this.sleepValue = sleepValue;    }    @Override    public String call() throws Exception {        Thread.sleep(sleepValue);        return "return " + userName;    }}public class RunTest {    public static void main(String[] args){        MyCallable callable1 = new MyCallable("username1", 5000);        MyCallable callable2 = new MyCallable("username2", 4000);        MyCallable callable3 = new MyCallable("username3", 3000);        MyCallable callable4 = new MyCallable("username4", 2000);        MyCallable callable5 = new MyCallable("username5", 1000);        List<Callable> callableList = new ArrayList<Callable>();        callableList.add(callable1);        callableList.add(callable2);        callableList.add(callable3);        callableList.add(callable4);        callableList.add(callable5);        ThreadPoolExecutor executor = new ThreadPoolExecutor(5, 10, 5, TimeUnit.SECONDS, new LinkedBlockingDeque<>());        ExecutorCompletionService<Object> csRef = new ExecutorCompletionService<>(executor);        for(int i=0;i<5;i++){            csRef.submit(callableList.get(i));        }        try {            for(int i=0;i<5;i++){                System.out.println("wait for print " + (i+1) + " result");                System.out.println(csRef.take().get());            }        } catch (InterruptedException e) {            e.printStackTrace();        } catch (ExecutionException e) {            e.printStackTrace();        }        executor.shutdown();    }结果wait for print 1 resultreturn username5wait for print 2 resultreturn username4wait for print 3 resultreturn username3wait for print 4 resultreturn username2wait for print 5 resultreturn username1

使用CompletionService接口后,哪个任务先执行完,哪个任务的返回值就先打印。如果当前任务没有执行完,则csRef.take().get()方法还是阻塞。

take()

方法take()取得最先完成任务的Future对象,谁执行时间最短谁最先返回。

public class RunTest {    public static void main(String[] args){        ExecutorService executorService = Executors.newCachedThreadPool();        ExecutorCompletionService<String> csRef = new ExecutorCompletionService<String>(executorService);        for(int i=0;i<10;i++){            csRef.submit(new Callable<String>() {                @Override                public String call() throws Exception {                    long sleepValue = (int)(Math.random()*1000);                    System.out.println("sleep="+sleepValue+" "+ Thread.currentThread().getName());                    Thread.sleep(sleepValue);                    return "test " + Thread.currentThread().getName();                }            });        }        try {            for(int i=0;i<10;i++){                System.out.println(csRef.take().get());            }        } catch (InterruptedException e) {            e.printStackTrace();        } catch (ExecutionException e) {            e.printStackTrace();        }    }}

poll()

poll()获取并移除已完成任务的Future,非阻塞,不存在已完成任务则返回null

public class RunTest {    public static void main(String[] args) throws ExecutionException, InterruptedException {        ExecutorService executorService = Executors.newCachedThreadPool();        ExecutorCompletionService<String> csRef = new ExecutorCompletionService<String>(executorService);        for (int i = 0; i < 10; i++) {            csRef.submit(new Callable<String>() {                @Override                public String call() throws Exception {                    Thread.sleep(3000);                    return "test " + Thread.currentThread().getName();                }            });        }        System.out.println(csRef.poll());        Thread.sleep(4000);        System.out.println(csRef.poll().get());    }}结果nulltest pool-1-thread-1

统计程序运行时间

long start = System.currentTimeMillis();long end = System.currentTimeMillis();(end-start)/1000

poll()超时处理

poll(long timeout, TimeUnit unit)

public class RunTest {    public static void main(String[] args) throws ExecutionException, InterruptedException {        ExecutorService executorService = Executors.newCachedThreadPool();        ExecutorCompletionService<String> csRef = new ExecutorCompletionService<String>(executorService);        MyCallable callableA = new MyCallable("username1",2000);        MyCallable callableB = new MyCallable("username2",2000);        long start = System.currentTimeMillis();        csRef.submit(callableA);        csRef.submit(callableB);        for(int i=0;i<2;i++){            System.out.println("zzz "+ csRef.poll(6,TimeUnit.SECONDS).get());        }        System.out.println("main end!");        long end = System.currentTimeMillis();        // 关闭线程池确定任务结束        executorService.shutdown();        executorService.awaitTermination(Integer.MAX_VALUE, TimeUnit.SECONDS);        System.out.println("time is " + (end-start)/1000);    }}结果zzz return username2zzz return username1main end!time is 2

poll(long timeout,TimeUnit unit)方法是阻塞的

public class RunTest {    public static void main(String[] args) throws ExecutionException, InterruptedException {        ExecutorService executorService = Executors.newCachedThreadPool();        ExecutorCompletionService<String> csRef = new ExecutorCompletionService<String>(executorService);        MyCallable callableA = new MyCallable("username1",4000);        MyCallable callableB = new MyCallable("username2",4000);        long start = System.currentTimeMillis();        csRef.submit(callableA);        csRef.submit(callableB);        // poll方法在接下来的6秒之内会轮询,直到有数据返回继续向下执行,或者超时向下执行        System.out.println("zzz "+ csRef.poll(6,TimeUnit.SECONDS).get());        // 从结果可以看出,poll(long timeout,TimeUnit unit)是方法阻塞的        System.out.println("hello hello ");        System.out.println("zzz "+ csRef.poll(2,TimeUnit.SECONDS).get());        long end = System.currentTimeMillis();        executorService.shutdown();        executorService.awaitTermination(Integer.MAX_VALUE, TimeUnit.SECONDS);        System.out.println("main end!");        System.out.println("time is " + (end-start)/1000);    }}结果zzz return username1hello hello zzz return username2main end!time is 4

某一个任务失败不会影响这个任务执行之前成功的任务,却会导致之后的任务失败。

Future submit(Runnable task,V result)方法

public class RunTest {    public static void main(String[] args) throws ExecutionException, InterruptedException {        ExecutorService executorService = Executors.newCachedThreadPool();        ExecutorCompletionService<UserInfo> csRef = new ExecutorCompletionService<UserInfo>(executorService);        UserInfo userInfo = new UserInfo();        userInfo.setUsername("one");        System.out.println("before " + userInfo.getUsername());        // 注:UserIno类的run方法会抛出异常        MyRunnable myRunnable = new MyRunnable(userInfo);        Future<UserInfo> future = csRef.submit(myRunnable, userInfo);        System.out.println("after " + future.get().getUsername());    }}结果before onerunning ...after usernameValue
0 0
原创粉丝点击