spring注解@Async线程池（2）

问题

有些时候，我们需要等所有线程执行完后，再执行后面的业务代码。解决方案可以有：

方案一：使用信号量 CountDownLatch
1.自定义线程池配置(Bean定义)

@Bean("markUserThread")
public ThreadPoolTaskExecutor executeMarkUserThread() {
        ThreadPoolTaskExecutor poolExecutor = new ThreadPoolTaskExecutor();
        poolExecutor.setCorePoolSize(5);
        poolExecutor.setMaxPoolSize(60);
        poolExecutor.setQueueCapacity(5000);
        poolExecutor.setKeepAliveSeconds(10);
        poolExecutor.setThreadNamePrefix("markUserThread-");
        poolExecutor.setWaitForTasksToCompleteOnShutdown(true);
        poolExecutor.setRejectedExecutionHandler(new ThreadPoolExecutor.CallerRunsPolicy());
        poolExecutor.setAwaitTerminationSeconds(60);
        poolExecutor.initialize();
        return poolExecutor;
    }

定义了一个核心线程数为5，最大线程数60的线程池。

2.线程池Bean注入

@Resource(name = "markUserThread")
private ThreadPoolTaskExecutor poolTaskExecutor;

3.线程池使用

// 对大集合（任务）进行切分
List<List<Object>> partitionList = Lists.partition(dataList, 100);
        // 计算要运行的线程数
        int threadNum = partitionList.size();
        threadNum = Math.min(threadNum, 60);
         // 初始化信号量
        CountDownLatch latch = new CountDownLatch(threadNum);
        partitionList.forEach(list -> {
            poolTaskExecutor.submit(() ->                   
                // 业务代码
            );
            // 信号量减1
            latch.countDown();
        });
        try {
            // 等待所有线程执行完成
            latch.await();
            log.info("usePoolTaskExecutorMarkUserTag end->{}", threadNum);
        } catch (Exception e) {
            throw e;
        }

方案二：使用CompletableFuture
1.自定义线程池（同上）
2.注解@Async引入

@Async("markUserThread")
public CompletableFuture<String> dataListHandler(List<Integer> dataList) {
        dataList.forEach(rs -> {
            log.info("dataListHandler->{}", rs);
        });
        return CompletableFuture.completedFuture("success");
    }

3.业务调用（⚠️必须在其他类文件中调用加了Async注解的方法）

// 定义Future对象
List<CompletableFuture<?>> futures = Lists.newArrayList();
// 对大集合（任务）进行切分
List<List<Integer>> result = Lists.partition(elements, 100);
result.forEach(res -> {             
futures.add(asyncCommand.dataListHandler(res));
});            
// 等待所有线程执行完成
CompletableFuture.allOf(futures.toArray(new CompletableFuture<?>[0])).join();
log.info("task over!!!");