Redis批量设置Key过期

Redis批量¹设置Key过期，一般有两种方式：lua脚本和pipeline²。

1. lua脚本

-- expire_keys.lua 存放在classpath的script目录中
for i=1, ARGV[1], 1
do
    redis.call("PEXPIRE", KEYS[i], ARGV[2]);
end

//过期时间,pexpire的过期时间为ms单位，如果是秒请用expire
String EXPIRE_ACCOUNT_FEERATE = getExpireTime();

//redis key
List<String> redisKeys = getRedisKeys();

//加载脚本
ScriptSource scriptSource = new ResourceScriptSource(new ClassPathResource("script/expire_keys.lua"));
RedisScript redisScript = new DefaultRedisScript();
((DefaultRedisScript) redisScript).setScriptSource(scriptSource);

redisTemplate.execute(redisScript, redisKeys, String.valueOf(redisKeys.size()), EXPIRE_ACCOUNT_FEERATE);

2. pipeline³⁴

/** 通过Spring data redis来实现**/
//过期时间,pexpire的过期时间为ms单位，如果是秒请用expire
final long EXPIRE_ACCOUNT_FEERATE = getExpireTime();
//redis key
final Set<String> redisKeys = getRedisKeys();
StringRedisSerializer stringRedisSerializer = new StringRedisSerializer();
//redisTemplate.executePipelined适用于需要返回值
redisTemplate.execute(new RedisCallback<Object>() {
    @Override
    public Object doInRedis(RedisConnection connection) throws DataAccessException {
        for (String redisKey : redisKeys) {
            byte[] rawKeys = stringRedisSerializer.serialize(redisKey);
            //pexpire的过期时间为ms单位，如果是秒请用expire
            connection.pExpire(rawKeys, EXPIRE_ACCOUNT_FEERATE);
        }

        return null;
    }
});

参考资料：

---

1. Using pipelining to speedup Redis queries ↩︎

2. why do not Redis provide a batch set expire time operation on multiple keys? ↩︎

3. Spring data redis doc pipeline ↩︎

4. spring-data-redis 使用pipeline批量设置过期时间的bug ↩︎