Tip: after the article is written, the directory can be generated automatically. Please refer to the help document on the right for how to generate it
preface
Today, let's introduce the data structures of redis and their usage scenarios. My level is limited. If you are misled, welcome to learn together and make progress together!
1, 8 data types of redis?
There are 8 data types in redis, including 5 basic data types and 3 special data types
Basic data types: String, list (set), set (set), Zset (ordered set), hash (hash)
Special data types: geospatial (geographical location), hyperlog (advantage: small memory consumption), bitmap (bitmap)
1. string type
String type can store string, number and binary (bit). Common commands are as follows:
set key value ex 10 nx yes setnx key value；expire key 10 Integration of. Represents 10 second expiration. incr key Automatic pair key of value Add 1, value Must be numeric type decr key incrby key num Automatic pair key of value conduct, value+num Assigned to value Operation of, decrby key num
2. hash type
string is the most commonly used in hash, and the structure of hash is in the format of key file value. Common commands are as follows:
hset Deposit one key filed value Hash structure of hsetnx Deposit one key field value ，if key If it already exists, the operation fails and does not exist,success hget Gets the specified key hmset Batch deposit key field hmget Batch acquisition key field hdel Delete the specified key field hincrby yes key field To add or subtract the value of
3. list type
Common commands are as follows:
lpush key value [value ...] to key Put an element on the left of the list key, key New if none exists
rpush key value [value ...] to key Put an element on the right of the list key, key New if none exists
lpop key from key An element pops up at the leftmost end of the list key
rpop key from key An element pops up at the right end of the list key(delete from list)
lrange key start stop Get list key from start Subscript to stop Subscript element lrange key 0 -1 whole
blpop key[key...] timeout Blocked slave key An element pops up at the leftmost end of the list key. If there is no element in the list key,
Blocked waiting{timeout}Seconds, if{timeout}=0，Always blocked
brpop key[key...] timeout Blocked slave key An element pops up at the rightmost end of the list key. If there is no element in the list key,
Blocked waiting{timeout}Seconds, if{timeout}=0，Always blocked
4. set type
Common commands are as follows:
sadd key member [member...] Forward set key key Store elements in, do key If it does not exist, it will be added
srem key member [member...] From set key key Delete element from (srem sremove Delete meaning)
smembers key Get collection key key All elements in
scard key Get collection key key Number of elements
sismember key member judge member Whether the element exists in the collection key key in
srandmember key [count] From set key key Selected from{count}Element, not removed from the collection key
spop key [count] From set key key Selected from{count}Element and removed from the collection key
count Is a number, representing from key Delete so many elements from(Random deletion). redis3.2 Later versions support this command
SET If possible, it can also:
Intersection operation:
sinter key [key...]
sinterstore destination key [key...]
Union operation:
sunion key [key...]
sunionstore destination key [key...]
Difference set operation:
sdiff key [key...]
sdiffstore destination key [key...]
Suppose there are three set
set1 a,b,c
set2 b,c,d
set3 c,d,e
be
sinter set1 set2 set3 c
sinterstore destination set1 set2 set3 : take set1,2,3 Union of set enter key Value call destination inside
sunioin set1 set2 set3 a,b,c,d,e
sdiff set3 set2 set1 e (This has a reference, set3 and set2 By comparison, yes e yes set2 Not in. With set1 comparison, set1 either e，So it is e)
5. zset type command
zadd key score element [...] Forward ordered set key key Store elements in, score Is the score, element Is an element. if key If it does not exist, create a new one zrem key element [...] Forward ordered set key key Delete element from zscore key element Get ordered collection key key in element Elemental score value zincrby key increment element Ordered set key Medium element Element proceed score Value operation, increment The value of can be positive or negative. if key If it does not exist, create a new one, element If the element does not exist, add it score operation zcard key Get ordered collection key Number of elements in zrange key start stop [withscores] Get ordered set in positive order key from start Subscript to stop Subscript element. Plus the back withscores The corresponding score will also be printed. If it is not added, only the open element will be printed zrevrange key start stop [withscores] Get ordered set in reverse order key from start Subscript to stop Subscript element Set operation zunionstore destkey numkeys key [key..] Union calculation (Very common) zinterstore destkey numkeys key [key..] Intersection calculation (Not commonly used)
6,geospatial & hyperloglog & Bitmap
seeing the name of a thing one thinks of its function. geospatial is geographical location, hyperlog is log, and Bitmap is Bitmap. One of the characteristics of hyperloglog is that it occupies little memory.
2, Usage scenario
1. string type
- Data caching to speed up response time and improve user experience.
- Distributed lock (refer to the following "advantages of string over hash 3")
string and hash are almost the most commonly used, so I won't introduce them here
2. hash type
- Shopping cart function
Add hincrby {userid}: shoppingcart {goodsid} {count}
Query hget {userid}: shoppingcart
It means that each user has his own shopping cart (so the shopping cart is regarded as a key), and each commodity in the shopping cart is regarded as a field of hash structure; The quantity of each commodity is regarded as the value of the hash structure. If the user selects 2 for the number of pieces of the same commodity, add 1 to the value. Because it is the user id and commodity id, you can get the commodity unit price and commodity name. - Cache of menu data
For example, your menus include ICs KM system and ICs customer system. Each system has its own type and value. For example, the channels of the ics-km system include official website, Alipay, WeChat, many spells, and small programs. You can use hash structure key-field-value structure, key is ics-km, field is Alipay, value is 0. The code is as follows (example):
hset ics-km-channel Alipay 0 Wechat 1
3. list type
-
Message queuing can be implemented. Because it has a command called blpop key timeout. Blocked wait timeout seconds. The provider can be used to insert messages into the queue, and the consumer can block the consumption of messages from the queue. Although there is no sequential and persistent solution like the special message queue, this function can also be realized (it needs to be limited from other places)
-
You can implement the latest message of the list of concerns. For example, you pay attention to Zhang San and Li Si. When you read 2021.8.3, you should push the content of 2021.8.3; When you look at 2021.8.4, you should push the message of 2021.8.4. So how to achieve the latest news at the front? You can use the rpush key v of redis to push the latest message to the front.
4. set type
- Like a lucky draw. Microblog bloggers said that if you forward my microblog, I will draw a lottery at random. (or swiping a gift lottery or something). For example, whoever forwards your microblog, you add the user id to the key. Then the second person forwards it and adds the user id of the second person to the key. Finally, when it's time to draw, it depends on your rules. If a person can only win one prize, you can spop key 2. (assuming that there are two people in the second prize, two people will be randomly selected to win the prize, and they will be moved out of the key after winning the prize). If you can repeat the middle, you can srandmember key 3 and draw 3 winners at random. After winning, the winners are still in the key and can still participate in the next lucky draw.
- Can do some praise, sign in, my collection of knowledge... And so on
For example, if user 1001 likes a post with ID 8001, you can: SADD like_8001 1001 The user cancelled the like again: SREM like_8001 1001 Check whether user 1001 likes it SISMEMBER like_8001 1001 Get users who like post 8001 SMEMBERS like_8001 Total number of users who get likes of post 8001 SCARD like_8001
- Tiktok can be used as user recommendation, product recommendation, etc. (for example, people who may be concerned with shaking).
Zhang San(People concerned are): zhangSub --> {li,wang,zhao}
Li Si: liSub --> {zhang,wang,zhao,tian}
Wang Wu: wangSub --> {zhang,li,zhao,tian}
Zhang San opens Li Si's home page
Common concerns of Zhang San and Li Si: SINTER zhangSub liSub --> {wang,zhao}
Zhang San is also concerned about him(Mutual attention) SISMEMBER wangSub li,SISMEMBER zhaoSub li,.. (Traverse all the followers of Zhang San to see if there is Zhang San in the list of followers)
People Zhang San may know SDIFF liSub zhangSub --> {zhang,li} // Because Zhang San enters Li Si's home page and recommends people he may know to Zhang San, he recommends some of Li Si's concerns to Zhang San, but not Zhang San's
5. zset type
- One day ranking list (news hot list, knowledge hot list, etc.). For example, you can create a new key for zset:
zadd hostNewId_date 1 newId / / create a new hot news id_ key of date 1 news id
zincrby hostNewId_date 1 newId / / add 1 each time you click
zrevrange hostNewId_ Date 0 15 WithCores / / get the top 15 on the hot list and display the values. - Weekly list, monthly list, annual list, etc. For example, you use the knowledge Id_ The date is used as the key of zset of redis, and the combined key names of zunionstore are 7 key1... Key when making weekly list statistics
7 (the zunionstore is fixed, the merged keys are random, 7 is the number of keys, followed by the value of each key) count the keys for 7 days, and then use zrevrange key 0 10 to get the first 10 data. Similarly, you can get the monthly list and annual list.
3, Other
1. What are the advantages of hash over string?
1. Information can be gathered together for easy management
2. To some extent, misoperation can be avoided and key conflict can be reduced
3, Reduce memory / io/cpu consumption (this means that redis is aimed at scanning, retrieving, ttl expiration, etc. for the key. Although there are multiple field s in the key, there is still only one key. If a string is used, a large number of keys will be generated, and a scan of a large number of keys will be generated. If a key is used, the Ke of the outermost layer will be reduced Y, cpu consumption. For example, key conflicts, key management, etc. one key must consume less resources than multiple keys)
2. What are the advantages of string over hash?
1. The field of hash has no expired function. It can only set the expiration time of the key, and cannot set the expiration time for a single field (so the hash of redis cannot be used as a distributed lock, but only a string)
2. The hash structure does not support binary commands, while the string type supports (setbit, getbit, bitcount...)
3. In a redis cluster, you cannot use hash to distribute data. In redis3 In the custer cluster of 0, the key value is hashed and the module is taken (the hash nest is used to calculate the crc16 hash function based on the key, and then the module is taken with HA Xi nest, which is managed based on different hardware). When the key is the same, the location must be the same, that is, it must be on a physical machine, Storage that cannot be distributed. For string type, the key is different, while for hash type, the key is the same, and the fields in the key are different. However, when clustering, the operation is based on the key, not the field
summary
Here is just a brief introduction to the use scenario of redis. I will share with you the underlying principle and specific implementation of redis in the future.