List集合分批多线程处理，同时控制最大并发

业务中，要实现数据日终同步，采用将同步文件中的数据封装成List集合分批处理加多线程的方式，根据数据量动态设置线程数，同时控制最大并发数量（业务中有IO操作，避免过大并发导致堵塞），实现效率提高

//最大线程数控制
private static int MAX_THREADS= 5;
//跑批分页大小
private static int EXPIRED_PAGE_SIZE = 30;

private void dataHandler(List<SyncFileDto> list) {
    //处理数据数量
    int listSize = list.size();
    //线程数
    int runSize;
    if (listSize % EXPIRED_PAGE_SIZE == 0) {
        runSize = (listSize / EXPIRED_PAGE_SIZE);
    } else {
        runSize = (listSize / EXPIRED_PAGE_SIZE) + 1;
    }
    ThreadPoolExecutor executor = new ScheduledThreadPoolExecutor(runSize);
    CountDownLatch countDownLatch = new CountDownLatch(runSize);
    //最大并发线程数控制
    final Semaphore semaphore = new Semaphore(MAX_THREADS);
    List handleList = null;
    for (int i = 0; i < runSize; i++) {
        if ((i + 1) == runSize) {
            int startIndex = i * EXPIRED_PAGE_SIZE;
            int endIndex = list.size();
            handleList = list.subList(startIndex, endIndex);
        } else {
            int startIndex = i * EXPIRED_PAGE_SIZE;
            int endIndex = (i + 1) * EXPIRED_PAGE_SIZE;
            handleList = list.subList(startIndex, endIndex);
        }
        SyncTask task = new SyncTask(handleList, countDownLatch, semaphore);
        executor.execute(task);
    }
    try {
        countDownLatch.await();
    } catch (InterruptedException e) {
        e.printStackTrace();
    } finally{
    executor.shutdown();
    }
}

class SyncTask implements Runnable {
    private List<SyncFileDto> list;
    private CountDownLatch countDownLatch;
    private Semaphore semaphore;

    public SyncSyncTask(List<SyncFileDto> list, CountDownLatch countDownLatch, Semaphore semaphore) {
        this.list = list;
        this.countDownLatch = countDownLatch;
        this.semaphore = semaphore;
    }

    @Override
    public void run() {
        if (!CollectionUtils.isEmpty(list)) {
            try {
                semaphore.acquire();
                list.stream().forEach(fileDto -> {
                    //业务处理
                });

            } catch (InterruptedException e) {
                e.printStackTrace();
            } finally {
                semaphore.release();
            }

        }
        //线程任务完成
        countDownLatch.countDown();
    }
}

上面是通过手动数据分片，CountDownLatch计数器闭锁和Semaphore限流的方式进行的并发控制，后期改造时发现逻辑较复杂，因此改变线程池的类型，创建可控制的线程池ThreadPoolExecutor（该线程池也是ScheduledThreadPoolExecutor的父类），自定义其属性实现跑批线程池线程数量及并发量可控。

ExecutorService fixedThreadPool = new ThreadPoolExecutor(INIT_NTHREADS, INIT_NTHREADS, 10, TimeUnit.SECONDS, new ArrayBlockingQueue<Runnable>(1000),
                new ThreadPoolExecutor.CallerRunsPolicy());

该线程池的使用我在后面博客【地址】 中有介绍，可移步阅读。