一文了解：redis基础类型

redis特点

开源的，bsd许可高级的key-value存储系统
可以用来存储字符串，哈希结构，链表，集合

安装

windows：

mac\linux：

redis 数据类型

redis支持五种数据类型：string（字符串），hash（哈希），list（列表），set（集合）及zset（有序集合）。

string（字符串）

字符串类型是redis的最基本数据结构。
字符串类型的值实际可以为字符串，数字，二进制，但是值最大不能超过512m。

key	value
hello	world
counter	1
bits	10000100
json	{"id":1,"name":"xiaocai","age":18}

string命令

设置

set key value [ex seconds] [px milliseconds] [nx|xx]

get key

> set name xiaocai
ok
> get name
"xiaocai"

过期

setex key seconds value

> set name1 redis1 ex 10 #10秒后过期
ok
> get name1
(nil)

> setex name2 10 redis2 #10秒后过期
ok
> get name2
"redis2"
> get name2
(nil)

> set name3 redis3 px 10 #10豪秒后过期
ok
> get name3
(nil)

不存在才能设置成功或者必须存在才能设置成功

> set hello world nx #不存在才能设置成功
ok
> set hello w nx #存在就设置失败
(nil)
> get hello
"world"

>set hello w xx  #存在才能设置成功
ok
> set world hello xx #不存在就设置失败
(nil)
> get hello
"w"
> get world
(nil)

批量设置

mset key value [key value ...]
mget key [key ...]


> mset name1 redis1 name2 redis2
ok

> mget name1 name2
1) "redis1"
2) "redis2"

> mget name1 name2 name3
1) "redis1"
2) "redis2"
3) (nil)

计数

incr key

incrby key increment

> set age 18  #value只能为整数
ok
> incr age
(integer) 19
> incrby age -5
(integer) 14
> incrby age 10
(integer) 24

删除

del key [key ...]

> del age
(integer) 1
> get age
(nil)

内部编码

1. int 8个字节的长整型
2. embstr 小于等于39个字节的字符串
3. raw 大于39个字节的字符串

> set port 6379
ok
> object encoding port
"int"

> set hello world
ok
> object encoding hello 
"embstr"

> set longstring abcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyz
ok
> object encoding hello
"raw"
> strlen longstring
(integer) 52

list（列表）

list类型是用来存储多个有序的字符串。每列字符串称之为元素。一个list的最大存储为2^32-1个元素。可以对列表进行双端插入和弹出，也可以指定索引下标获取元素。

list命令

头部和尾部添加元素

lpush key value [value ...]

rpush key value [value ...]

lrange key start stop

> lpush letter "a"
(integer) 1
> lpush letter "b"
(integer) 2
> lpush letter "c"
(integer) 3
> lrange letter 0 -1
1) "c"
2) "b"
3) "a"

> rpush letter "a"
(integer) 4
> rpush letter "b"
(integer) 5
> rpush letter "c"
> lrange letter 0 -1
1) "c"
2) "b"
3) "a"
4) "a"
5) "b"
6) "c"

头部和尾部弹出元素

lpop key

rpop key

> lpop letter
"c"
> lpop letter
"b"
> lrange letter 0 -1
1) "a"
2) "a"
3) "b"
4) "c"

> rpop letter
"c"
> rpop letter
"b"
> lrange letter 0 -1
1) "a"
2) "a"

索引操作
索引需要对全部list进行遍历，性能会随着元素个数的增大而变差

lrange key start stop

lindex key index

ltrim key start stop

len key

> rpush letter b c
(integer) 4
> lrange letter 0 -1
1) "a"
2) "a"
3) "b"
4) "c"
> lindex letter 2
"b"
> ltrim letter 0 -2
ok
> lrange letter 0 -1
1) "a"
2) "a"
3) "b"
> llen letter
(integer) 3

插入

insert key before|after pivot value

> linsert letter before b c
(integer) 4
> linsert letter after a d
(integer) 5
> lrange letter 0 -1
1) "a"
2) "d"
3) "a"
4) "c"
5) "b"

修改

set key index value

> lset letter 2 b
ok
> lrange letter 0 -1
1) "a"
2) "d"
3) "b"
4) "c"
5) "b"

内部编码

1. ziplist（压缩列表）：小于3.2版本，当元素个数小于list-max-ziplist-entries配置（默认512个），同时每个元素的值长度都小于list-max-ziplist-value配置（默认64字节）
2. linkedlist（链表）：小于3.2版本，不满足ziplist的条件
3. quicklist：redis 3.2版本，以一个ziplist为节点的linkedlist

> object encoding letter
"quicklist"

hash (哈希)

hash是一个string类型的field和value的映射表。 它适合用于存储对象，它是无序的，不能使用索引操作。

hash命令

设置

hset key field value

> hset user:1 name zhangsan
(integer) 1
> hset user:1 age 18
(integer) 1

获取和获取所有的field-value

hget key field

hgetall key

> hget user:1 name
"zhangsan"
> hgetall user:1
1) "name"
2) "zhangsan"
3) "age"
4) "18"

删除

hdel key field [field ...]

> hdel user:1 age
(integer) 1
> hdel user:2 age
(integer) 0

长度

hlen key

> hlen user:1
(integer) 1

批量设置

hmset key field value [field value ...]

> hmset user:2 name lisi age 23
ok
> hmget user:2 name age
1) "lisi"
2) "23"
> hlen user:1
(integer) 1
> hlen user:2
(integer) 2

是否存在

hexists key field

> hexists user:2 name
(integer) 1
> hexists user:2 city
(integer) 0

所有的field和所有的value

hkeys key

hvals key

> hkeys user:1
1) "name"
> hkeys user:2
1) "name"
2) "age"
> hvals user:1
1) "zhangsan"
> hvals user:2
1) "lisi"

内部编码

1. ziplist（压缩列表）：当元素个数小于hash-max-ziplist-entries配置（默认512个）和每个元素大小小于hash-max-ziplist-value配置（默认64字节）时
2. hashtable（哈希表）：不满足ziplist条件时

> object encoding user:1
"ziplist"

修改配置文件hash-max-ziplist-entries为5

> hmset test t1 v1 t2 v2 t3 v3 t4 v4 t5 v5 t6 v6
ok
> object encoding test
"hashtable"

set（集合）

用来保存多个字符串元素，无序的，不能有重复元素，不能使用索引下标获取元素。一个集合可以存储2^32-1个元素。

set命令

增加

add key member [member ...]

> sadd user:1:tag it music news
(integer) 3
> sadd user:1:tag bike news
(integer) 1

删除

srem key member [member ...]

> srem user:1:tag bike
(integer) 1

个数

scard key

> scard user:1:tag
(integer) 3

是否存在

sismember key member

> sismember user:1:tag bike
(integer) 0
> sismember user:1:tag it
(integer) 1

随机返回指定个数

srandmember key [count]

> srandmember user:1:tag
"news"
> srandmember user:1:tag 3
1) "news"
2) "music"
3) "it"

随机弹出

spop key [count]

> spop user:1:tag
"news"
> srandmember user:1:tag 3
1) "music"
2) "it"

所有个数

smembers key

> smembers user:1:tag
1) "music"
2) "it"

交集

sinter key [key ...]

> sinter user:1:tag user:2:tag
1) "music"
2) "it"

并集

sunion key [key ...]

> sunion user:1:tag user:2:tag
1) "music"
2) "java"
3) "run"
4) "it"

差集

sdiff key [key ...]

> sdiff user:1:tag user:2:tag
(empty list or set)
> sadd user:1:tag sleep
(integer) 1
> sdiff user:1:tag user:2:tag
1) "sleep"
> sdiff user:2:tag user:1:tag
1) "java"
2) "run"

内部编码

1. intset（整数集合）：元素都是整数和元素个数小于set-max-intset-entries配置（默认512个）时
2. hashtable（哈希表）：不满足intset时

> object encoding user:1:tag
"hashtable"

> sadd numbers 1 2 3 4 5
(integer) 5
> object encoding numbers
"intset"

zset（有序集合）

zset保证了元素不能重复，每个元素都有一个分数（score）作为排序的依据。

zset命令

添加

zadd key [nx|xx] [ch] [incr] score member [score member ...]

> zadd books 8.2 "redis in action"
(integer) 1
> zadd books 9.3 "effective java: second edition : java"
(integer) 1
> zadd books 9.1 "think in java"
(integer) 1
> zadd books 9.3 "python cookbook" 9.0 "effective python"
(integer) 2

个数

zcard key

> zcard books
(integer) 5

升序返回范围的成员

zrange key start stop [withscores]

> zrange books 0 -1
1) "redis in action"
2) "effective python"
3) "think in java"
4) "effective java: second edition : java"
5) "python cookbook"

> zrange books 2 5
1) "think in java"
2) "effective java: second edition : java"
3) "python cookbook"

升序返回成员时带上分数

> zrange books 0 -1 withscores
 1) "redis in action"
 2) "8.1999999999999993" 
 3) "effective python"
 4) "9"
 5) "think in java"
 6) "9.0999999999999996"
 7) "effective java: second edition : java"
 8) "9.3000000000000007"
 9) "python cookbook"
10) "9.3000000000000007"

> zrange books 2 5 withscores
1) "think in java"
2) "9.0999999999999996"
3) "effective java: second edition : java"
4) "9.3000000000000007"
5) "python cookbook"
6) "9.3000000000000007"

1. 精度问题：内部 score 使用 double 类型进行存储，所以存在小数点精度问题
2. withscores:带上分数

降序

zrevrange key start stop [withscores]

> zrevrange books 0 -1 withscores
 1) "python cookbook"
 2) "9.3000000000000007"
 3) "effective java: second edition : java"
 4) "9.3000000000000007"
 5) "think in java"
 6) "9.0999999999999996"
 7) "effective python"
 8) "9"
 9) "redis in action"
10) "8.1999999999999993"

指定value的score

zscore key member

> zscore books "think in java"
"9.0999999999999996"

根据score的数值区间升序

zrangebyscore key min max [withscores] [limit offset count]

> zrangebyscore books 0 9.1 withscores
1) "redis in action"
2) "8.1999999999999993"
3) "effective python"
4) "9"
5) "think in java"
6) "9.0999999999999996"

根据score的数值区间降序

zrevrangebyscore key max min [withscores] [limit offset count]

> zrevrangebyscore books 9.1 0  withscores
1) "think in java"
2) "9.0999999999999996"
3) "effective python"
4) "9"
5) "redis in action"
6) "8.1999999999999993"

根据score的数值降序输出所有元素

> zrevrangebyscore books +inf -inf  withscores
 1) "python cookbook"
 2) "9.3000000000000007"
 3) "effective java: second edition : java"
 4) "9.3000000000000007"
 5) "think in java"
 6) "9.0999999999999996"
 7) "effective python"
 8) "9"
 9) "redis in action"
10) "8.1999999999999993"

1. +inf 正无穷
2. -inf 负无穷

删除

zrem key member [member ...]

> zrem books "effective java: second edition : java"
(integer) 1

增加分数

zincrby key increment member

> zincrby books 2  "redis in action"
"10.199999999999999"
> zincrby books -1  "redis in action"
"9.1999999999999993"

交集

zinterstore destination numkeys key [key ...] [weights weight] [aggregate

> zadd textbooks 8 "chinese" 9 "english" 9.5 "mathematics"
(integer) 3
> zadd textbooks 9.2 "think in java"
(integer) 1

> zrange books 0 -1
1) "redis in action"
2) "effective python"
3) "think in java"
4) "python cookbook"
> zrange textbooks 0 -1
1) "chinese"
2) "english"
3) "think in java"
4) "mathematics"

> zinterstore newbooks 2 books textbooks
(integer) 1
> zrange newbooks 0 -1
1) "think in java"

并集

zunionstore destination numkeys key [key ...] [weights weight] [aggregate

> zunionstore newunionbooks 2 books textbooks
(integer) 7
> zrange newunionbooks 0 -1
1) "chinese"
2) "redis in action"
3) "effective python"
4) "english"
5) "python cookbook"
6) "mathematics"
7) "think in java"

内部编码

1. ziplist（压缩列表）：元素个数小于zset-max-ziplist-entries配置（默认128个）和元素长度小于zset-max-ziplist-value配置（默认64b）时
2. skiplist（跳跃表）：不满足ziplist时

> object encoding books
"ziplist"

结语

本人深知水平有限，欢迎指正本文错误之处。

参考资料

《redis 实战》

《redis 开发与运维》

《redis 深度历险：核心原理与应用实践》