Redis使用

基本操作

General

# 返回给定模式的keys
KEYS patter
KEYS * # 返回全部
KEYS set* # 返回set开头的keys
EXISTS key
TYPE key
DEL key

String

SET key value
GET key
# Set Extend Time
SETEX key seconds value
# Set When Key Not Exist
SETNX key value

Hash

HSET key field value
HGET key field
HDEL key field
# Get All Fields
HKEYS key
# Get All Values
HVALS key

flowchart LR
    key[key]
    item[
    field1: value1
    field2: value2
    ]
    key --> item

List

LPUSH key value1 value2
# Get Key From Start To Stop
LRANGE key start stop
# Right POP
RPOP key
# List Length
LLEN key

典型场景

栈

订阅消息

队列，先来后到

• 如微信、微博订阅消息
• 阻塞队列B<L|R>POP，队列为空就等待

Set

SADD key mem1 mem2
SMEMBERS key
# Set Size
SCARD key
SINTER key1 key2
SUNION key1 key2
# Delete
SREM key mem1 mem2

典型场景

抽奖

SADD lottery {user_id}
SMEMBERS lottery
# 开奖（适用于单个奖品）
SRANDMEMBER lottery {drawing_count}
# 开奖并删除（适用于多项奖品，不重复得奖）
SPOP lottery {drawing_count}

点赞、收藏、标签

# 点赞
SADD like:{msg_id} {user_id}
# 取消
SREM like:{msg_id} {user_id}
# 用户是否点赞
SISMEMBER like:{msg_id} {user_id}
# 点赞用户列表
SMEMBERS like:{msg_id}
# 点赞用户数
SCARD like:{msg_id}

关注、商品筛选

利用集合特性运算

• 如共同关注、推荐关注
• 商品筛选

Sorted Set / ZSet

ZADD key score1 mem1 score2 mem2
# Show List
ZRANGE key start stop (WITHSCORES)
# Increse Member
ZINCRBY key increment member
ZREM key mem1 mem2

典型场景

日/月/年热点排行榜

# 记录浏览量
ZINCRBY hotNews:{date} 1 {news_id}
# Top 10
ZREVRANGE hotNews:{date} 0 9 WITHSCORES
# Recent 7 days
ZUNIONSTORE hotNews:{start_date}-{end_date} 0 9 WITHSCORES

Redis for Java

• Jedis (Official Recommand)
• Lettuce
• Spring Data Redis

Redis Data Redis

pom.xml

<!-- Redis -->
<dependency>
    <groupId>org.springframwork.boot</groupId>
    <artifactId>spring-boot-starter-data-redis</artifactId>
</dependency>

application.yml

spring:
  redis:
    host: localhost
    port: 6379
    password: yourPassword

@Configuration
@Slf4j
public class RedisConfiguration {
    @Bean
    public RedisTemplate redisTemplate(RedisConnectionFactory redisConnectionFactory) {
        log.info("开始创建Redis模板对象...");
        RedisTemplate redisTemplate = new RedisTemplate();
        // Set Redis Connection Factory Object
        redisTemplate.setConnectionFactory(redisConnectionFactory);
        // Set Key Serializer
        redisTemplate.setKeySerializer(new StringRedisSerializer());
        return redisTemplate;
    }
}

持久化

RDB

# redis-cli
save
# 将会生成一个dump.rdb
bgsave
# 后台保存

AOF

保存执行的命令为日志。每次重启加载所有命令。

# redis.windows.conf
appendonly yes
appendsync always/everysec/no
auto-aof-rewrite-percentage 100
auto-aof-rewrite-min-size 64mb

# redis-cli
bgrewriteaof # 重新编排命令，让重启时执行更高效

事务

# 开始事务
multi

set key value
# ... your command
# 中途取消
discard

# 执行
exec 

乐观锁

乐观锁：不认为别人会来抢占资源，所以会直接对数据进行操作，在操作时验证是否资源已被占用。

乐观锁会比较数据是否和原数据一致，一致，说明没有人抢占资源，可以修改。

watch key
# 通过版本号，而不通过值来判断
unwatch key

典型场景

对象缓存

SET user:1 <json>
MSET user:1:name the_name user1:age the_age

分布式锁

SETNX product:10003 true ex 10 nx
# success return 1, or fail return 0

... # your operation

# release the lock
DEL product:10003

计数器

INCR article:readcount:{article_id}
GET article:readcount:{article_id}

Web集群共享session

Spring Session + Redis

分布式系统全局序列号

INCRBY order_id 1000

Scan 流式遍历

SCAN cursor <KEY RegularExpression> <Count>

cursor相当于一个下标，每次查询，scan会给出下一次开始的cursor，由此实现分批查询。

Redis Lua

Lua脚本可以一次性原子地执行多条redis语句。减少网络开销、替代Redis事务。

A Redis script is transactional by defination, so everything your can do with a Redis transaction, you can also do with a script, and usually the script will be both simpler and faster.

EVAL script numkeys key [key...] arg [arg...]

numkeys: 键名参数个数
接着可以以1为基址访问keys，如KEYS[1]

$ eval "return {KEYS[1], KEYS[2], ARGV[1], ARGV[2]}" 2 key1 key2 first second

key1
key2
first
second

Jedis使用lua脚本

jedis.set("prod_stock_10016", "15");
String script = """
    local count = redis.call('get', KEYS[1])
    local a = tonumber(count)
    local b = tonumber(ARGV[1])
    if a >= b then
        redis.call('set', KEYS[1], a - b)
        return 1 
    end
    return 0
"""

Object obj = jedis.eval(script, Arrays.asList("prod_stock_10016"), Arrays.asList("10"));