代码之家 › 专栏 › 技术社区 › Questioner

由多个线程并行计算哈希,并将输出添加到ArrayList<String>

concurrency arraylist multithreading java

Questioner · 技术社区 · 3 年前

为了计算 String (基于SHA-256),然后将所有输出插入 ArrayList<String> :

        ArrayList<Thread> threadList = new ArrayList<Thread>();
        ArrayList<String> threadListStr = new ArrayList<String>();
        int threadNumber = 100;
        for (int i = 0; i < threadNumber; i++) {
            String tId = String.valueOf(i);
            Thread thr = new Thread(() -> {
                threadListStr.add(calculateHash(tId));
            });
            threadList.add(thr);
        }

        // START the threads
        for (int i = 0; i < threadNumber; i++) {
            threadList.get(i).start();
        }
        // STOP the threads
        for (int i = 0; i < threadNumber; i++) {
            threadList.get(i).interrupt();
        }

        System.out.println("Size of ArrayList<String> is: " + threadListStr.size());
        System.out.println("Size of ArrayList<Thread> is: " + threadList.size());
        
        /////////////////////
        
        public static String calculateHash(String tId) {
        String tIdStr = org.apache.commons.codec.digest.DigestUtils.sha256Hex(tId);
        return tIdStr;
        }

但是,ArrayList并不完整,正如您在运行代码5次后看到的,每次ArrayList都有不同的大小(尽管ArrayList线程列表总是完整的,因为线程数是100)

//1th run
Size of ArrayList<String> is: 60
Size of ArrayList<Thread> is: 100

//2nd run
Size of ArrayList<String> is: 30
Size of ArrayList<Thread> is: 100

//3rd run
Size of ArrayList<String> is: 10
Size of ArrayList<Thread> is: 100

//4th run
Size of ArrayList<String> is: 61
Size of ArrayList<Thread> is: 100

//5th
Size of ArrayList<String> is: 69
Size of ArrayList<Thread> is: 100

应该如何修改代码,以便 ArrayList<字符串> 商店 全部的 输出 彻底地 ?

伊迪特: 我对代码做了如下更改,但输出是一样的。

        ArrayList<Thread> threadList = new ArrayList<Thread>();
        //ArrayList<String> threadListStr = new ArrayList<String>();
        List<String> threadListStrSync = Collections.synchronizedList(new ArrayList<>());
        int threadNumber = 100;
        for (int i = 0; i < threadNumber; i++) {
            String tId = String.valueOf(i);
            Thread thr = new Thread(() -> {
                threadListStrSync.add(calculateHash(tId));
            });
            threadList.add(thr);
        }

        // START the threads
        for (int i = 0; i < threadNumber; i++) {
            threadList.get(i).start();
        }
        // STOP the threads
        for (int i = 0; i < threadNumber; i++) {
            threadList.get(i).interrupt();
        }

        System.out.println("Size of ArrayList<String> is: " + threadListStrSync.size());
        System.out.println("Size of ArrayList<Thread> is: " + threadList.size());

注: 我评论道 interrupt(); 但产出仍然是一样的。

0 回复 | 直到 3 年前

Hulk 3 年前

有很多问题

使用threadsafe集合或手动同步访问——一个简单的选择是用 Collections.synchronizedList()
interrupt() 如果不需要,线程将在到达任务末尾时终止 run() -方法
在打印结果之前,需要等待所有线程终止——为此,请调用 join() 而不是 中断()

Dave S 3 年前

您有两个问题:1)可能有一些线程在执行过程中始终无法将其哈希ID添加到集合中,2)哈希ID的集合正被多个线程访问,因此您应该使用线程安全的集合。

    ArrayList<Thread> threadList = new ArrayList<Thread>();
    Collection<String> threadListStr = Collections.synchronizedCollection( new ArrayList<String>() );
    int threadNumber = 100;
    for (int i = 0; i < threadNumber; i++) {
        String tId = String.valueOf(i);
        Thread thr = new Thread(() -> {
            threadListStr.add(calculateHash(tId));
        });
        threadList.add(thr);
    }

    // START the threads
    for (int i = 0; i < threadNumber; i++) {
        threadList.get(i).start();
    }
    // STOP the threads
    for (int i = 0; i < threadNumber; i++) {
       try {
           threadList.get(i).join();
       } catch( InterruptedException exc ) {
           // handle interrupted exception
       }
    }

    System.out.println("Size of ArrayList<String> is: " + threadListStr.size());
    System.out.println("Size of ArrayList<Thread> is: " + threadList.size());