一文了解：Redis基礎型別

Redis特點

開源的，BSD許可高階的key-value儲存系統
可以用來儲存字串，雜湊結構，連結串列，集合

安裝

windows：https://github.com/microsoftarchive/redis/releases

mac\linux：http://www.redis.cn/

Redis 資料型別

Redis支援五種資料型別：string（字串），hash（雜湊），list（列表），set（集合）及zset（有序集合）。

string（字串）

字串型別是Redis的最基本資料結構。
字串型別的值實際可以為字串，數字，二進位制，但是值最大不能超過512M。

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

string命令

設定

set key value [EX seconds] [PX milliseconds] [NX|XX]

get key

> set name xiaocai
OK
> get name
"xiaocai"

過期

setex key seconds value

> set name1 redis1 ex 10 #10秒後過期
OK
> get name1
(nil)

> setex name2 10 redis2 #10秒後過期
OK
> get name2
"redis2"
> get name2
(nil)

> set name3 redis3 px 10 #10豪秒後過期
OK
> get name3
(nil)
 

不存在才能設定成功或者必須存在才能設定成功

> set hello world nx #不存在才能設定成功
OK
> set hello w nx #存在就設定失敗
(nil)
> get hello
"world"

>set hello w xx  #存在才能設定成功
OK
> set world hello xx #不存在就設定失敗
(nil)
> get hello
"w"
> get world
(nil)

批量設定

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)
 

計數

incr key

incrby key increment

> set age 18  #value只能為整數
OK
> incr age
(integer) 19
> incrby age -5
(integer) 14
> incrby age 10
(integer) 24

刪除

del key [key ...]

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

內部編碼

1. int 8個位元組的長整型
2. embstr 小於等於39個位元組的字串
3. raw 大於39個位元組的字串

> 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型別是用來儲存多個有序的字串。每列字串稱之為元素。一個list的最大儲存為2^32-1個元素。可以對列表進行雙端插入和彈出，也可以指定索引下標獲取元素。

list命令

頭部和尾部新增元素

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"

頭部和尾部彈出元素

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
索引需要對全部list進行遍歷，效能會隨著元素個數的增大而變差

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"

內部編碼

1. ziplist（壓縮列表）：小於3.2版本，當元素個數小於list-max-ziplist-entries配置（預設512個），同時每個元素的值長度都小於list-max-ziplist-value配置（預設64位元組）
2. linkedlist（連結串列）：小於3.2版本，不滿足ziplist的條件
3. quicklist：Redis 3.2版本，以一個ziplist為節點的linkedlist

> object encoding letter
"quicklist"

hash (雜湊)

hash是一個string型別的field和value的對映表。 它適合用於儲存物件，它是無序的，不能使用索引操作。

hash命令

設定

hset key field value

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

獲取和獲取所有的field-value

hget key field

hgetall key

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

刪除

hdel key field [field ...]

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

長度

hlen key

> hlen user:1
(integer) 1

批量設定

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"

內部編碼

1. ziplist（壓縮列表）：當元素個數小於hash-max-ziplist-entries配置（預設512個）和每個元素大小小於hash-max-ziplist-value配置（預設64位元組）時
2. hashtable（雜湊表）：不滿足ziplist條件時

> object encoding user:1
"ziplist"

修改配置檔案hash-max-ziplist-entries為5

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

set（集合）

用來儲存多個字串元素，無序的，不能有重複元素，不能使用索引下標獲取元素。一個集合可以儲存2^32-1個元素。

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

刪除

srem key member [member ...]

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

個數

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

隨機返回指定個數

srandmember key [count]

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

隨機彈出

spop key [count]

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

所有個數

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"

並集

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"

內部編碼

1. intset（整數集合）：元素都是整數和元素個數小於set-max-intset-entries配置（預設512個）時
2. hashtable（雜湊表）：不滿足intset時

> 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保證了元素不能重複，每個元素都有一個分數（score）作為排序的依據。

zset命令

新增

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

個數

zcard key

> zcard books
(integer) 5

升序返回範圍的成員

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"

升序返回成員時帶上分數

> 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. 精度問題：內部 score 使用 double 型別進行儲存，所以存在小數點精度問題
2. withscores:帶上分數

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"

根據score的數值區間升序

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"

根據score的數值區間降序

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"

根據score的數值降序輸出所有元素

> 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 正無窮
2. -inf 負無窮

刪除

zrem key member [member ...]

> zrem books "Effective Java: Second Edition : Java"
(integer) 1

增加分數

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"

並集

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"

內部編碼

1. ziplist（壓縮列表）：元素個數小於zset-max-ziplist-entries配置（預設128個）和元素長度小於zset-max-ziplist-value配置（預設64B）時
2. skiplist（跳躍表）：不滿足ziplist時

> object encoding books
"ziplist"

結語

本人深知水平有限，歡迎指正本文錯誤之處。

參考資料

《Redis 實戰》

《Redis 開發與運維》

《Redis 深度歷險：核心原理與應用實踐》

相關推薦