java并发:ExecutorServiec中的CompletionService和invokeAll的使用

package com;import java.util.ArrayList;import java.util.List;import java.util.Random;import java.util.concurrent.BlockingQueue;import java.util.concurrent.Callable;import java.util.concurrent.CompletionService;import java.util.concurrent.ExecutionException;import java.util.concurrent.Executor;import java.util.concurrent.ExecutorCompletionService;import java.util.concurrent.ExecutorService;import java.util.concurrent.Executors;import java.util.concurrent.Future;import java.util.concurrent.LinkedBlockingDeque;import java.util.concurrent.TimeUnit;import java.util.concurrent.TimeoutException;/** * java 并发 * http://blog.csdn.net/lmj623565791/article/details/27250059 *  * 带返回结果的批量任务执行 * CompletionService *  * ExectorService.invokeAll * */public class CompletionServiceExecutorServiceDemo {public static void main(String[] args) throws Exception {//doCompletionService();doExecutorServiceInvokeAll();}/** 演示CompletionService  * @throws ExecutionException  * @throws InterruptedException *  * CompletionService将Executor（线程池）和BlockingQueue（阻塞队列）结合 * 在一起，同时使用Callable作为任务的基本单元，整个过程就是生产者不断把Callable * 任务放入阻塞对了，Executor作为消费者不断把任务取出来执行，并返回结果； *  */public static void doCompletionService() throws InterruptedException, ExecutionException{// 初始化定义ExecutorService executor = Executors.newFixedThreadPool(10);// 容量为10的阻塞队列 BlockingQueue<Future<Integer>> completionQueue  = new LinkedBlockingDeque<Future<Integer>>(10);CompletionService<Integer> service = new ExecutorCompletionService<Integer>(executor, completionQueue);/* 开始执行任务 * 模拟产生了10个任务 */for(int i = 0 ; i < 20 ; i ++){Future<Integer> submit = service.submit(new Callable<Integer>() { //submit 返回的引用,可以用来获取对应的返回的值 .// Callable 相当于 Runnable@Overridepublic Integer call() throws Exception { //任务在这里面执行int sleep = new Random().nextInt(10000);Thread.sleep(sleep);System.out.println(Thread.currentThread().getName() + " 休息了" + sleep + "毫秒");return sleep;}});//System.out.println("submit:" + submit.get());}//读取返回的结果 , 谁有结果返回, 就是返回谁的结果 , 在这里可以做一个无限的循环作为一个遍历for(int i = 0 ; i < 10 ; i ++){Future<Integer> take = service.take();System.out.println("take:" + take.get());//future.get(timeout, unit);可以设置超时的世家}//情况线程池executor.shutdown();}/** 演示ExecutorService  *  * ExecutorService的invokeAll方法也能批量执行任务，并批量返回结果，但是呢， * 有个我觉得很致命的缺点，必须等待所有的任务执行完成后统一返回，一方面内存持有的 * 时间长；另一方面响应性也有一定的影响，毕竟大家都喜欢看看刷刷的执行结果输出，而不是苦苦的等待； * @throws InterruptedException  * @throws ExecutionException  * @throws TimeoutException  * */public static void doExecutorServiceInvokeAll() throws InterruptedException, ExecutionException, TimeoutException{ExecutorService mThreadPool = Executors.newFixedThreadPool(10);List<Callable<Integer>> tasks = new ArrayList<Callable<Integer>>();for(int i = 0 ; i < 20 ; i++){Callable<Integer> callable = new Callable<Integer>() {@Overridepublic Integer call() throws Exception {int sleep = new Random().nextInt(2000);Thread.sleep(sleep);System.out.println(Thread.currentThread().getName() + "休息了:" + sleep);return sleep;}};tasks.add(callable);}long startTime = System.currentTimeMillis();// 执行并返回了结果List<Future<Integer>> invokeAll = mThreadPool.invokeAll(tasks);long allTime = System.currentTimeMillis() - startTime;System.out.println("执行了:" + allTime);//将所有的结果答应出来for(Future<Integer> future : invokeAll){System.out.println(future.get()); //future.get(timeout, unit);在该方式下似乎并没有用}mThreadPool.shutdown();}}