java多线程并发 Java多线程并发执行demo代码实例

package java8test;

import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.BlockingQueue;
import java.util.concurrent.Callable;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Future;
import java.util.concurrent.LinkedBlockingQueue;

/**
 * @param <H> 为被处理的数据类型
 * @param <T>返回数据类型
 * 知识点1：X,T为泛型，为什么要用泛型，泛型和Object的区别请看：https://www.cnblogs.com/xiaoxiong2015/p/12705815.html
 */
public abstract class MultiThread<X, T> {

  public static int i = 0;

  // 知识点2：线程池:https://www.cnblogs.com/xiaoxiong2015/p/12706153.html
  private final ExecutorService exec; // 线程池

  // 知识点3：@author Doung Lea 队列：https://www.cnblogs.com/xiaoxiong2015/p/12825636.html
  private final BlockingQueue<Future<T>> queue = new LinkedBlockingQueue<>();

  // 知识点4：计数器，还是并发包大神 @author Doug Lea 编写。是一个原子安全的计数器，可以利用它实现发令枪
  private final CountDownLatch startLock = new CountDownLatch(1); // 启动门，当所有线程就绪时调用countDown

  private final CountDownLatch endLock; // 结束门

  private final List<X> listData;// 被处理的数据

  /**
   * @param list list.size()为多少个线程处理，list里面的H为被处理的数据
   */
  public MultiThread(List<X> list) {
    if (list != null && list.size() > 0) {
      this.listData = list;
      exec = Executors.newFixedThreadPool(Runtime.getRuntime().availableProcessors()); // 创建线程池，线程池共有nThread个线程
      endLock = new CountDownLatch(list.size()); // 设置结束门计数器，当一个线程结束时调用countDown
    } else {
      listData = null;
      exec = null;
      endLock = null;
    }
  }

  /**
   * 
   * @return 获取每个线程处理结速的数组
   * @throws InterruptedException
   * @throws ExecutionException
   */
  public List<T> getResult() throws InterruptedException, ExecutionException {

    List<T> resultList = new ArrayList<>();
    if (listData != null && listData.size() > 0) {
      int nThread = listData.size(); // 线程数量
      for (int i = 0; i < nThread; i++) {
        X data = listData.get(i);
        Future<T> future = exec.submit(new Task(i, data) {

          @Override
          public T execute(int currentThread, X data) {

            return outExecute(currentThread, data);
          }
        }); // 将任务提交到线程池
        queue.add(future); // 将Future实例添加至队列
      }
      startLock.countDown(); // 所有任务添加完毕，启动门计数器减1，这时计数器为0，所有添加的任务开始执行
      endLock.await(); // 主线程阻塞，直到所有线程执行完成
      for (Future<T> future : queue) {
        resultList.add(future.get());
      }
      exec.shutdown(); // 关闭线程池
    }
    return resultList;
  }

  /**
   * 每一个线程执行的功能，需要调用者来实现
   * @param currentThread 线程号
   * @param data 每个线程被处理的数据
   * @return T返回对象
   */
  public abstract T outExecute(int currentThread, X data);

  /**
   * 线程类
   */
  private abstract class Task implements Callable<T> {

    private int currentThread;// 当前线程号

    private X data;

    public Task(int currentThread, X data) {
      this.currentThread = currentThread;
      this.data = data;
    }

    @Override
    public T call() throws Exception {

      // startLock.await(); // 线程启动后调用await，当前线程阻塞，只有启动门计数器为0时当前线程才会往下执行
      T t = null;
      try {
        t = execute(currentThread, data);
      } finally {
        endLock.countDown(); // 线程执行完毕，结束门计数器减1
      }
      return t;
    }

    /**
     * 每一个线程执行的功能
     * @param currentThread 线程号
     * @param data 每个线程被处理的数据
     * @return T返回对象
     */
    public abstract T execute(int currentThread, X data);
  }
}

package java8test;
public class ResultVo {
  int i;
  public ResultVo(int i) {
    this.i = i;
  }
  public ResultVo() {
    // TODO Auto-generated constructor stub
  }
}

package java8test;

public class ParamVo {

  private int i;

  ParamVo(int i) {
    this.i = i;
  }

  public int getI() {

    return i;
  }

  @Override
  public String toString() {

    return String.valueOf(i) + " " + hashCode();
  }
}

package java8test;

import java.util.ArrayList;
import java.util.List;

public class Test {

  public static void main(String[] args) {

    try {
      List<ParamVo> splitList = new ArrayList<ParamVo>();
      for (int i = 0; i < 100; i++) {
        splitList.add(new ParamVo(i));
      }
      List<ParamVo> splitList1 = new ArrayList<ParamVo>();
      for (int i = 200; i < 300; i++) {
        splitList1.add(new ParamVo(i));
      }
      MultiThread<ParamVo, ResultVo> multiThread = new MultiThread<ParamVo, ResultVo>(splitList) {

        @Override
        public ResultVo outExecute(int currentThread, ParamVo data) {

          System.out.println("当前线程名称：" + Thread.currentThread().getName() + "当前线程号=" + currentThread
              + " data=" + data);
          i--;
          return new ResultVo(data.getI());
        }
      };

      MultiThread<ParamVo, ResultVo> multiThread1 = new MultiThread<ParamVo, ResultVo>(splitList1) {

        @Override
        public ResultVo outExecute(int currentThread, ParamVo data) {

          System.out.println("当前线程名称：" + Thread.currentThread().getName() + "当前线程号=" + currentThread
              + " data=" + data);
          i--;
          return new ResultVo(data.getI());
        }
      };
      List<ResultVo> list = multiThread.getResult();
      List<ResultVo> list1 = multiThread1.getResult();
      // 获取每一批次处理结果
      System.out.println("获取处理结果........................");
      for (ResultVo vo : list) {
        System.out.println(vo.i);
      }
      System.out.println("获取1处理结果........................");
      for (ResultVo vo : list1) {
        System.out.println(vo.i);
      }
    } catch (Exception e) {
      e.printStackTrace();
    }
  }
}