Learn more about redis' eight data types and application scenario analysis

Introduce

Redis There are several types of data ？（ Note that the data type is not a data structure ）

There are eight kinds altogether ,String、Hash、Set、List、Zset、Hyperloglog、Geo、Streams.

1、 Why put data in memory ？

1. Memory is faster ,10W QPS
2. Reduce computing time , Reduce database pressure

2、 If you use the data structure of memory as a cache , Why not HashMap perhaps Memcached？

1. Richer data types
2. Support for multiple programming languages
3. Rich in functions ： Persistence mechanism 、 Memory retirement strategy 、 Business 、 Publish subscribe 、pipeline、LUA
4. Support clusters 、 Distributed

3、Memcached and redis What's the difference ？

Memcached Can only store KV、 There is no persistence mechanism 、 Master slave replication is not supported 、 It's multi-threaded .

Reids The default is 16 Databases （0-15）, Can be found in the configuration file redis.conf Revision in China .

databases 16

Redis Key The maximum length limit is 512M, Value limits differ , Some are limited by length , Some are limited by number .

String

Introduce

The most basic and commonly used data type is String.get and set The order is String Operation command of ,Redis The string is called Binary safe string （Binary-safe strings）.

String There are three types of storage ,INT( Integers )、float( Single-precision floating-point )、string( character string ).

Operation command

Here are some of its operating commands

#  Store value （ If to the same key set Multiple times will directly overwrite the old value ）
set jack 2673
#  Value 
get jack
#  Look at all the keys 
keys *
#  Get the total number of keys （ Large amount of production environment data , Use with caution ）
dbsize
#  Check if the key exists 
exists jack
#  Delete key 
del jack tonny
#  Rename key 
rename jack tonny
#  View type 
type jack 
#  Get the specified range of characters 
getrange jack 0 1
#  Get value length 
strlen jack
#  String append content 
append jack good
#  Set multiple values （ The batch operation , Atomicity ）
mset jack 2673 tonny 2674
#  Get multiple values 
mget jack tonny
#  Set the value , If key There is , No success 
setnx jack shuaige
#  Based on this, distributed locking is realized 
set key value [expiration EX seconds|PX milliseconds][NX|XX]
# ( Integers ) Value increment （ If the value does not exist, you will get 1）
incr jack
incrby jack 100
# ( Integers ) The value is decreasing 
decr jack
decrby jack 100
#  Floating point increments 
set mf 2.6
incrbyfloat mf 7.3

Application scenarios

1、 cache

String type , This is the most commonly used , You can cache some hot data , For example, home news , It can significantly improve the access speed of hot data , At the same time reduce DB pressure .

2、 Distributed data sharing

String type , because Redis It's a distributed, stand-alone service , It can be shared among multiple applications .

for example ： Distributed Session

<dependency>
  <groupId>org.springframework.session</groupId>
  <artifactId>spring-session-data-redis</artifactId>
</dependency>

3、 Distributed lock

Reference resources 《 This paper introduces several common writing methods of distributed lock 》

4、 overall situation ID

INT type ,INCRBY, Using atomicity

incrby userid 1000

（ The scenario of dividing tables and databases , Take one piece at a time ）

5、 Counter

INT type ,INCR Method

for example ： The amount of reading , Microblog likes , Allow a certain delay , Write... First Redis Then synchronize to the database regularly

6、 Current limiting

INT type ,INCR Method

With the visitor's IP And other information as key, Add one count at a time , If it exceeds the number of times, it will return false.

Hash

Introduce

Now there's one teacher surface

Suppose we still pass String Type storage , When storing, you should put Teacher Serialize entity classes , And then as a value Store only ; At the time of revision , Also need to put value Take it out first and deserialize it , For example, change the age to 21 year , And then serialize , Save it again , It's very complicated , Increased overhead .

We need to get it separately 、 Modify a value , Then we can go through key In a hierarchical way , The following table ：

But this way key It's going to be a lot ,key It's a long time , Occupancy space , Is there a better way , That's when we use our Hash type , The following two tables show ：

This also facilitates centralized management , The granularity of the partition is different , You can follow the actual scene ,key The expiration time of , Flexibility considers which storage method to choose .

Hash Used to store multiple unordered key value pairs , Maximum storage 2^32-1（40 Million or so ）.

advantage ：

1. Aggregate all related values into one Key in , Save memory space
2. Use only one Key, Reduce Key Conflict
3. When you need to get values in batches , Just use one command , Reduce memory /IO/CPU Consumption of

shortcoming ：

1. Field Expiration time cannot be set separately
2. We need to consider the distribution of data volume （field A lot of times , Cannot distribute to multiple nodes ）

Operation command

#  Set up 、 Batch setting value 
hset h1 f 6
hset h1 e 5
hmset h1 a 1 b 2 c 3 d 4
#  Value 
hget h1 a
#  Batch values 
hmget h1 a b c d
#  Get all field
hkeys h1
#  Get all field Value 
hvals h1
#  Return to the hash table , All fields and values 
hgetall h1
#  Delete field
hdel h1 a
#  Get the number of fields in the hash table 
hlen h1

Application scenarios

String What can be done ,Hash You can do it. . Add another scene , The shopping cart ：

key： user id;field： goods id;value： The number ;

+1：hincr;

-1：hdecr;

Delete ：hincrby key field -1;

Future generations ：hgetall;

Number of goods ：hlen;

List

Introduce

Store ordered strings （ From left to right ）, Elements can be repeated , Maximum storage 2^32-1（40 Million or so ）.

Let's show how to enter the queue by drawing a picture , Outgoing queue

Operation command

#  Push left 
lpush queue a
lpush queue b c
#  Push right 
rpush queue d e
#  The left removes and returns the first element of the list 
lpop queue
#  The right removes and returns the first element of the list 
rpop queue
#  Get the elements in the list by index 
lindex queue 0
#  Returns the elements in the specified interval in the list 
lrange queue 0 -1

Application scenarios

1、 list

For example, the user's message list 、 List of announcements on the website 、 Activity list 、 List of blog posts 、 Comment list, etc , adopt LRANGE Take out a page , Show in order .

2、 queue / Stack

List It can also be used as a queue in a distributed environment / Stack usage .

queue ： fifo ,rpush and blpop

Stack ： First in, then out ,rpush and brpop

Here are two blocked pop-up operations ：blpop/brpop, Timeout can be set （ Company ： second ）.

blpop：blpop key1 timeout, Move out and get the first element of the list ,  If there are no elements in the list, it will block the list until the wait timeout or pop-up elements are found .
brpop：brpop key1 timeout, Move out and get the last element of the list , If there are no elements in the list, it will block the list until the wait timeout or pop-up elements are found .

Set

Introduce

Set Storage String Unordered collection of type , Maximum storage 2^32-1（40 Million or so ）.

As shown in the figure below ：

Operation command

#  Add one or more elements 
sadd myset a b c d e f g
#  Get all the elements 
smembers myset
#  Number of statistical elements 
scard myset
#  Get an element at random 
srandmember myset
#  Pop an element at random 
spop myset
#  Remove one or more elements 
srem myset d e f
#  See if the element exists 
sismember myset a
#  Get the difference set 
sdiff set1 set2
#  Get intersection 
sinter set1 set2
#  Get Union 
sunion set1 set2

Application scenarios

1、 Luck draw

Random access to elements ：spop myset

2、 give the thumbs-up 、 Sign in 、 Clock in

Let's take Weibo as an example , Suppose this microblog ID yes t1001, user ID yes u6001,

use dianzan:t1001 To maintain the t1001 All the like users of this microblog .

I like this microblog ：sadd dianzan:t1001 u6001

Cancel likes ：srem dianzan:t1001 u6001

Do you like it ：sismember dianzan:t1001 u6001

All users who like ：smembers dianzan:t1001

Number of likes ：scard dianzan:t1001

Much simpler than relational databases .

3、 Product label

use tags:i8001 To maintain all labels on the product .

sadd tags:i8001 The picture is clear and delicate

sadd tags:i8001 True color clear screen

sadd tags:i8001 It's extremely smooth

4、 Product selection

Huawei P40 launched , Support national brands , Add to each label .

sadd brand:huawei p40

sadd os:android p40

sadd screensize:6.0-6.24 p40

Screening when buying , The brand is Huawei , The operating system is Android , The screen size is 6.0-6.24 Between , intersect ：

sinter brand:huawei os:android screensize:6.0-6.24

ZSet

Introduce

sorted set Store ordered elements . Every element has a score, according to score Sort from small to large .score Phase at the same time , according to key Of ASCII Sort code .

Operation command

#  Additive elements 
zadd myzset 10 java 20 php 30 ruby 40 cpp 50 python
#  Get all the elements 
zrange myset 0 -1 withscores
zrevrange myzset 0 -1 withscores
#  Get the element according to the score interval 
zrangebyscore myzset 20 30
#  Remove elements （ It can also be based on score rank Delete ）
zrem myzset php cpp
#  Number of statistical elements 
zcard myzset
#  The score increases 
zincrby myzset 5 python
#  According to the score min and max Number of Statistics 
zcount myzset 20 60
#  obtain python ranking 
zrank myzset python
#  Get element scores 
zscore myzset python

Application scenarios

Ranking List

It's today 2021 year 5 month 23 Number , To build a key by hotSearch:20210523 Of zset.

The news of Uncle sheep herding id yes n1234, Every click ：zincrby hotSearch:20210523 1 n1234

Get the top 10 hot search rankings ：zrevrange hotSearch:20210523 0 10 withscores

Yuanlao is an unparalleled scholar , All the way walk good , The children of the Chinese nation will never forget you ！

BitMaps

Introduce

BitMaps Is to define the position operation on the string type , One byte by 8 Binary bits make up . As shown in the figure below ：

m Of ASCII Code is 109, The corresponding binary data is 0110 1101

Operation command

#  Set string key by k1,value by mic
set k1 mic
#  take k1 Seventh place in , The result is 0
getbit k1 6
#  take k1 The eighth place is 0, At this time ASCII Code is 108, The corresponding letter is l
setbit k1 7 0
#  So the value is lic
get k1
#  In statistical binary 1 The number of , Is the total 12 individual 
bitcount k1
#  Get the first one 1 perhaps 0 The location of 
bitpos k1 1
bitpos k1 1

Application scenarios

• BITOP AND destkey key [key ...], To one or more key Seek logic and , And save the result to destkey.
• BITOP OR destkey key [key ...], To one or more key Seek logic or , And save the result to destkey.
• BITOP XOR destkey key [key ...], To one or more key Seek logical XOR , And save the result to destkey.
• BITOP NOT destkey key, For given key Logic is not , And save the result to destkey.

Online users for seven consecutive days

setbit firstday 0 1 // Set the first day uid yes 0 User login 
setbit firstday 1 0 // Set the first day uid yes 1 The user of is not logged in 
setbit firstday 2 1 // Set the first day uid yes 2 User login 
...
setbit secondday 0 0 // Set the next day uid yes 0 The user of is not logged in 
setbit secondday 1 1 // Set the next day uid yes 1 User login 
setbit secondday 2 1 // Set the next day uid yes 2 User login 
... // And so on 

So in the calculation, online users for seven consecutive days are ：

BITOP AND 7_both_online_users firstday secondday thirdday fourthday fifthday sixthday seventhday

You can also apply access statistics 、 Online user statistics and so on .

Hyperloglog

Hyperloglog Provides an imprecise cardinality statistical method , Used to count the number of non repeating elements in a collection , For example, Statistics website UV, Or applied daily life 、 Monthly living , There is a certain error .

stay Redis Implemented in Hyperloglog, It only needs 12k Memory can count 2^64 Data .

    public static void main(String[] args) {
        Jedis jedis = new Jedis("39.103.144.86", 6379);
        float size = 100000;
        for (int i = 0; i < size; i++) {
            jedis.pfadd("hll", "hll-" + i);
        }
        long total = jedis.pfcount("hll");
        System.out.println(String.format(" Number of Statistics : %s", total));
        System.out.println(String.format(" Accuracy rate : %s", (total / size)));
        System.out.println(String.format(" Error rate : %s", 1 - (total / size)));
        jedis.close();
    }

The source code in ：
com/xhj/jedis/HyperLogLogTest.java

Geo

Now there is a need , Get the radius 1 Stores within km , Then we should save the longitude and latitude of the store , If there is a database , A field stores longitude , A field is stored in a dimension , It is complicated to calculate the distance . Now let's go through Redis Of Geo Storage is very convenient .

#  Save latitude and longitude 
geoadd location 121.445	31.213 shanghai
#  Take latitude and longitude 
geopos location shanghai

Address location information can be added 、 Get address location information 、 Calculate the distance between two positions 、 Get the geographic location collection within the specified range, and so on . The source code in ：
com/xhj/jedis/GeoTest.java

    public static void main(String[] args) {
        Jedis jedis = new Jedis("39.103.144.86", 6379);
        Map<String, GeoCoordinate> geoMap = new HashMap<>();
        GeoCoordinate coordinate = new GeoCoordinate(121.445, 31.213);
        geoMap.put("shanghai", coordinate);
        jedis.geoadd("positions", geoMap);
        System.out.println(jedis.geopos("positions", "shanghai"));
        jedis.close();
    }

Streams

5.0 The type of data launched . Persistent message queue supporting multicast , It is used to implement the publish and subscribe function , Learn from it Kafka The design of the .

Summary of application scenarios

• cache — Speed up access to hot data
• Shared data — The problem of data storage and sharing
• overall situation ID— Distributed global ID The generation scheme of （ Sub database and sub table ）
• Distributed lock — Atomic operation guarantee of data sharing between processes
• Online user statistics and counting
• queue 、 Stack — Processes across queues / Stack
• Message queue — Asynchronous decoupling message mechanism
• Service registration and discovery —RPC Service coordination center of communication mechanism （Dubbo Support Redis）
• The shopping cart
• Sina user message timeline
• Draw logic （ gift 、 forward ）
• give the thumbs-up 、 Sign in 、 Clock in
• Product label
• user （ goods ） Focus on （ recommend ） Model
• E-commerce product screening
• Ranking List

Reproduced in ：
https://juejin.cn/post/6965870956103335949#heading-0