redis管道实现批量插入数据（lpush和hset举例）

项目背景：多线程执行文件ftp传输操作，成功一个redis记录++

文件数据量：百亿

线程数：500+

出现问题：并发执行incr redis数据，出现问题，redis的请求数量有限导致无法操作redis

示例一：list的leftpush操作

    改造前：单线程一个个排队插入数据

f.transferFileHistoryList.stream().forEach(x->{
    JSONObject file = new JSONObject();
    file.put("dataId", x.getDataId());
    file.put("fileId", x.getId());
    file.put("path", x.getPath());
    file.put("name", x.getName());
    RedisTaskExecutor.getExecutor("leftpush-%d").execute(() -> {
        redisTemplate.opsForList().leftPush(Constant.FILE_TRANSFER_QUEUE, JSON.toJSONString(file));
    });
});

改造后：

RedisSerializer<String> serializer = redisTemplate.getStringSerializer();
redisTemplate.executePipelined(new RedisCallback<String>() {
        @Override
        public String doInRedis(RedisConnection redisConnection) throws DataAccessException {
            for (int i = 0; i < 3000000; i++) {
                JSONObject file = new JSONObject();
                file.put("dataId", CommonUtil.newUUID());
                file.put("fileId", CommonUtil.newUUID());
                file.put("path", CommonUtil.newUUID());
                file.put("name", CommonUtil.newUUID());
                file.put("id", i);
                redisConnection.lPush(Constant.FILE_TRANSFER_QUEUE.getBytes(), JSON.toJSONString(file).getBytes());
            }
            return null;
        }
},serializer);

测试数据量：300万，耗时：70秒

优点：要么都成功，要么都失败

示例二：hash的hset操作

private void addHashRedis(TransferBean transferBean, FTPListAllFiles f) {
    // 记录成品库的redis
    String key = Constant.FILE_TRANSFER_PROGRESS + transferBean.getDataId();
    // 记录线程执行情况
    String threadKey = Constant.THREAD_PROGRESS + transferBean.getDataId();

    RedisSerializer<String> serializer = redisTemplate.getStringSerializer();
    redisTemplate.executePipelined(new RedisCallback<String>() {
        @Override
        public String doInRedis(RedisConnection redisConnection) throws DataAccessException {
            redisConnection.hSet(key.getBytes(), "total".getBytes(), (f.totalCount + "").getBytes());
            redisConnection.hSet(key.getBytes(), "success".getBytes(), "0".getBytes());
            redisConnection.hSet(key.getBytes(), "size".getBytes(), (f.totalSize + "").getBytes());

            redisConnection.hSet(threadKey.getBytes(), "total".getBytes(), (f.totalCount + "").getBytes());
            redisConnection.hSet(threadKey.getBytes(), "success".getBytes(), "0".getBytes());
            return null;
        }
    },serializer);
}

其它：还可以使用lua脚本方法执行