java多线程查询提高效率

最新推荐文章于 2024-09-30 09:20:12 发布

#老程

最新推荐文章于 2024-09-30 09:20:12 发布

阅读量381

点赞数

分类专栏： Java 文章标签： java

老程出品

本文链接：https://blog.csdn.net/LuckKing0925/article/details/107182937

版权

Java 专栏收录该内容

20 篇文章 1 订阅

订阅专栏

本文介绍了一种利用多线程加速数据处理的方法，通过将任务分配到多个线程中并行执行，显著提高了处理效率。文章详细展示了如何使用Java的CompletableFuture和ExecutorService来实现这一目标，并提供了一个具体的业务实现示例。

摘要由CSDN通过智能技术生成

工具

public <T> void accelerateThread(Consumer<List<T>> consumer,List<T> targetList,Integer threadSize) throws InstantiationException, IllegalAccessException, InterruptedException, ExecutionException {
		if (targetList == null || targetList.size() == 0) {
			return;
		}
		
		if (threadSize == null) {
			throw new RuntimeException("线程大小（threadSize）不能为空！");
		}
		
		if (threadSize > 100) {
			throw new RuntimeException("线程大小不能超过100！");
		}
		
		List<Future<?>> resultList = new ArrayList<Future<?>>();
		
		int listSize = targetList.size();
		Integer divideSize = targetList.size() / threadSize;
		
		if(divideSize.intValue() == 0) {
			divideSize = targetList.size();
			threadSize = 1;
		}
		
		CompletableFuture<Void> completableFuture = null;
		
		for (int i = 0; i < threadSize; i++) {
			List<T> list = (i == threadSize-1) ? targetList.subList(i* divideSize, listSize)
					: targetList.subList(i* divideSize, (i + 1) * divideSize);
	      	completableFuture = CompletableFuture.runAsync(()->consumer.accept(list),threadPoolTaskExecutor);
			
			resultList.add(completableFuture);
		}
		
        CompletableFuture.allOf(resultList.stream().toArray(CompletableFuture[]::new)).join();

	}

public static TaskExecutor getInstance() {
		if (taskExecutor == null) {
			taskExecutor = new TaskExecutor();
		}
		return taskExecutor;
	}

具体业务实现举例

/**
     * 制作多线程
     *      传来 models 用户集合;制作多线程分别对用户同时进行相同的操作，提高速度
     * @return
     * @throws BusinessException
     */
    public List<Model> getThreadList(List<Model> models) throws BusinessException {
        List<Model> threadList = Collections.synchronizedList(new ArrayList<>());
        Consumer<List<Model>> consumer = temps -> {
        //写业务操作
            for (Model model : temps) {
                Boolean flag = poService.isGoodModel(model.getNick());     //用户是好用户
                if (flag) {
                    threadList.add(model);
                }
            }
        };

        try {
            TaskExecutor.getInstance().accelerateThread(consumer, models, 100);
        } catch (InstantiationException e) {
            e.printStackTrace();
        } catch (IllegalAccessException e) {
            e.printStackTrace();
        } catch (InterruptedException e) {
            e.printStackTrace();
        } catch (ExecutionException e) {
            e.printStackTrace();
        }

        return threadList;
    }