程序中有个从字典表获取数据的service -- TDicdataServiceImpl。

考虑到频繁获取字典数据，后来做了本地缓存。实现方案是利用ScheduledThreadPoolExecutor#schedule 。 在频繁访问这个方法过程中，设定每10分钟清理内存数据。

package com.cn.yft.ora.service.impl;

import com.cn.yft.ora.dao.TDicdataDAO;
import com.cn.yft.ora.entity.TDicdata;
import com.cn.yft.ora.service.TDicdataService;
import com.yft.util.StringUtil;
import lombok.extern.slf4j.Slf4j;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Service;

import java.util.HashMap;
import java.util.List;
import java.util.concurrent.ScheduledThreadPoolExecutor;
import java.util.concurrent.TimeUnit;

@Slf4j
@Service
public class TDicdataServiceImpl implements TDicdataService {

    private static HashMap<String,List<TDicdata>> cacheByDicmemo = new HashMap<>();
    private static ScheduledThreadPoolExecutor executor = new ScheduledThreadPoolExecutor(1);

    @Autowired
    private TDicdataDAO tDicdataDAO;

    @SuppressWarnings("unchecked")
    @Override
    public List<TDicdata> selectByDicmemo(String dicmemo) {
        List<TDicdata> list = null;
        if(StringUtil.isEmpty(dicmemo)) return null;
        list = cacheByDicmemo.get(dicmemo);
        if(list == null || list.size() == 0){
            list = this.tDicdataDAO.selectByDicDesc(dicmemo);
            cacheByDicmemo.put(dicmemo, list);
        }
        if(executor.getActiveCount()+executor.getQueue().size()==0) {
            executor.schedule(this::cacheByDicmemoClean,10, TimeUnit.MINUTES);
        }
        return list;
    }

    @Override
//    @Scheduled(cron = "0 0/10 * * * ?")  注解@Scheduled对spring无效，需在spring.xml单独配置CronTriggerBean
    public void cacheByDicmemoClean() {
        if(!cacheByDicmemo.isEmpty()) {
            int size=cacheByDicmemo.size();
            cacheByDicmemo.clear();
            log.info("#字典缓存已清空---清空缓存---本地缓存--已清空本地内存数据{}条", size);
        }
    }
}

 

说来真巧。上面缓存优化是在2020年年底改造的。 而今适逢一年，旧历2021年的年底，使用Hutool-cache的LFUCache来再做一次改造升级。

package com.cn.yft.ora.service.impl;

import cn.hutool.cache.CacheUtil;
import cn.hutool.cache.impl.LFUCache;
import com.cn.yft.ora.dao.TDicdataDAO;
import com.cn.yft.ora.entity.TDicdata;
import com.cn.yft.ora.service.TDicdataService;
import com.yft.util.StringUtil;
import lombok.extern.slf4j.Slf4j;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Service;
import sun.reflect.generics.reflectiveObjects.NotImplementedException;

import java.util.HashMap;
import java.util.List;
import java.util.concurrent.ScheduledThreadPoolExecutor;
import java.util.concurrent.TimeUnit;

@Slf4j
@Service
public class TDicdataServiceImpl implements TDicdataService {

    // 缓存TTL：10分钟
    private static LFUCache<String,List<TDicdata>> cacheByDicmemo = CacheUtil.newLFUCache(16, TimeUnit.MINUTES.toMillis(10));

    @Autowired
    private TDicdataDAO tDicdataDAO;

    @SuppressWarnings("unchecked")
    @Override
    public List<TDicdata> selectByDicmemo(String dicmemo) {
        List<TDicdata> list = null;
        if(StringUtil.isEmpty(dicmemo)) return null;
        list = cacheByDicmemo.get(dicmemo);
        if(list == null || list.size() == 0){
            list = this.tDicdataDAO.selectByDicDesc(dicmemo);
            log.debug("获取key={},结果条数={}", dicmemo, list.size());
            cacheByDicmemo.put(dicmemo, list);
        }
        return list;
    }
    @Override
    public void cacheByDicmemoClean() {
        throw new NotImplementedException();
    }
}