java多线程使用方法之一

1. 首先引入spring自带的ThreadPoolTaskExecutor构建线程的方法：

@Autowired
private ThreadPoolTaskExecutor threadPoolTaskExecutor;

也可以自己写一个配置类配置线程池：

@EnableAsync
@Configuration
public class ThreadPoolTaskExecutorConfig {

    /**
     * 线程池维护线程的最少数量
     */
    private final static int CORE_POOL_SIZE = 20;
    /**
     * 线程池维护线程的最大数量(超过最大值，workQueue将拒绝执行任务)
     */
    private final static int MAX_POOL_SIZE = 20;
    /**f
     * 线程池空闲线程存活的时间
     */
    private final static int KEEP_ALIVE_SECONDS = 60;
    /**
     * 线程池被阻塞线程队列容量
     */
    private final static int BLOCKING_QUEUE_CAPACITY = 1200;

    @Bean
    public ThreadPoolTaskExecutor threadPoolTaskExecutor() {
        ThreadPoolTaskExecutor executor = new ThreadPoolTaskExecutor();
        executor.setCorePoolSize(CORE_POOL_SIZE);
        executor.setMaxPoolSize(MAX_POOL_SIZE);
        executor.setKeepAliveSeconds(KEEP_ALIVE_SECONDS);
        executor.setQueueCapacity(BLOCKING_QUEUE_CAPACITY);
        return executor;
    }
}

然后在使用@Qualifier() 注解使用自己指定的配置类：

@Autowired
@Qualifier("threadPoolTaskExecutor")
private ThreadPoolTaskExecutor threadPoolTaskExecutor;

2. 调用threadPoolTaskExecutor的api：

threadPoolTaskExecutor.execute(Runnable tack);

tack 就是你要处理的业务代码逻辑（封装在一个类），要实现Runnable 的run 方法。

以下例子：同步一批数据到ES

public class GoodsInfo580Task implements Runnable {

    private List<GoodsInfo580DTO> data;

    private GoodsInfo580Repository goodsInfo580Repository;

    public GoodsInfo580Task(){

    }

    public GoodsInfo580Task(List<GoodsInfo580DTO> data,GoodsInfo580Repository goodsInfo580Repository){
        this.data = data;
        this.goodsInfo580Repository = goodsInfo580Repository;
    }

    @Override
    public void run() {
        List<GoodsInfo580Document> dtos = new ArrayList<>(data.size());
        for (GoodsInfo580DTO dto : data) {
            if(Func.isNull(dto)){
                continue;
            }
            GoodsInfo580Document goodsInfo580Document = new GoodsInfo580Document();
            HealthGoodsInfoConstructHelper.setHealthGoodsInfo(dto,goodsInfo580Document);
            dtos.add(goodsInfo580Document);
        }
        try {
            log.info("插入ES数量：{}",dtos.size());
            goodsInfo580Repository.saveAll(dtos);
        } catch (Exception e) {
            log.error("全量同步580商品到ES异常：{},错误信息：{}", JSONObject.toJSONString(dtos),e);
        }
    }
}

3. 也可以不创建一个tack类，直接调用execute 方法
   例子：计算一批数据的商品价格

public void calculate580Pres() {
    //查询580处方药品信息
    List<PrescriptionDTO> prescriptionDTOList = this.baseMapper.selectPresInfo();
    if(Func.isEmpty(prescriptionDTOList)){
        log.error("查询580处方信息为空：{}",prescriptionDTOList);
        return;
    }
    //多线程执行计算订单总价
    threadPoolTaskExecutor.execute(()->{
        for (PrescriptionDTO prescriptionDTO : prescriptionDTOList) {
            // 判断是否存在刷取订单价格
            Object lock = redisUtils.get(RedisConstant.REDIS_580LOCK + prescriptionDTO.getId());
            if (Func.isNotEmpty(lock)){
                continue;
            }
            Integer storeId = prescriptionDTO.getStoreId();
            String drugRspList = prescriptionDTO.getDrugRspList();
            List<DrugRspListDTO> listDTOS = JSONArray.parseArray(drugRspList, DrugRspListDTO.class);
            double price = 0.00;
            try {
                // 药品品种
                int size = listDTOS.size();
                // 判断计算后的药品数量
                int sum = 0;
                //有多药需要累积计算
                for (DrugRspListDTO listDTO : listDTOS) {
                    if (listDTO.getAmout() > 0) {
                        GoodsUnionIdDTO byGoodsId = iSkuGoodsMapperClient.getByGoodsId(1000001L, listDTO.getCommodityCode());
                        log.info("商品参数：{}",byGoodsId);
                        String goodsId = byGoodsId.getGoodsId().toString();
                        BigDecimal priceByParams = iPriceDeptGoodsClient.getPriceByParams(1000001L, storeId.toString(), goodsId);
                        log.info(goodsId + "商品价格：{}",priceByParams);
                        if (Func.isNotEmpty(priceByParams) && priceByParams.compareTo(BigDecimal.ZERO) > 0){
                            BigDecimal count = BigDecimal.valueOf(priceByParams.doubleValue() * listDTO.getAmout()).setScale(2);
                            if (count.compareTo(BigDecimal.ZERO)>0){
                                ++sum;
                            }
                            price = price + count.doubleValue();
                            log.info("计算处方单总价：{}", prescriptionDTO.getId() + "：" +price);
                        }
                    }
                }
                if (price > 0 && Func.equals(size,sum)){
                    this.baseMapper.updatePriceById(Double.toString(price),prescriptionDTO.getId());
                    redisUtils.set(RedisConstant.REDIS_580LOCK + prescriptionDTO.getId(),"1",60*3);
                }else {
                    this.baseMapper.updatePriceTimeById(prescriptionDTO.getId());
                }
            } catch (Exception e) {
                this.baseMapper.updatePriceTimeById(prescriptionDTO.getId());
                log.error("计算处方单总价失败：{}",prescriptionDTO.getId() + "：" +price);
            }
        }
    });
}