Redis getting started complete tutorial: hash description

Almost all programming languages provide hashes （hash） type , Their name may be ha
Greek 、 Dictionaries 、 Associative array . stay Redis in , Hash type means that the key itself is a key value pair
structure , Form like value={ {field1,value1},...{fieldN,valueN}},Redis Key value pairs and
The relationship between hash type and hash type can be shown in Figure 2-14 To express .

The mapping relationship in a hash type is called field-value, Notice the value Refer to field Corresponding
Value , The key of the value is not , Please note that value Role in different contexts .

2.3.1　 command
（1） Set the value
hset key field value
The following is user：1 Add a pair of field-value：
127.0.0.1:6379> hset user:1 name tom
(integer) 1
If the setting is successful, it will return 1, On the contrary, it will return 0. Besides Redis Provides hsetnx command , it
Our relationship is like set and setnx command , It's just that the scope changes from bond to field.
（2） Get value
hget key field
for example , The following operation gets user：1 Of name Domain （ attribute ） Corresponding value ：
127.0.0.1:6379> hget user:1 name
"tom"
If the key or field non-existent , Returns the nil：
127.0.0.1:6379> hget user:2 name
(nil)
127.0.0.1:6379> hget user:1 age
(nil)
（3） Delete field

hdel key field [field ...]
hdel Will delete one or more field, The result is successfully deleted field The number of , for example ：
127.0.0.1:6379> hdel user:1 name
(integer) 1
127.0.0.1:6379> hdel user:1 age
(integer) 0
（4） Calculation field Number
hlen key
for example user：1 Yes 3 individual field：
127.0.0.1:6379> hset user:1 name tom
(integer) 1
127.0.0.1:6379> hset user:1 age 23
(integer) 1
127.0.0.1:6379> hset user:1 city tianjin
(integer) 1
127.0.0.1:6379> hlen user:1
(integer) 3
（5） Batch set or get field-value
hmget key field [field ...]
hmset key field value [field value ...]
hmset and hmget They are batch setting and obtaining field-value,hmset The required parameter is key
And many pairs field-value,hmget The required parameter is key And multiple field. for example ：
127.0.0.1:6379> hmset user:1 name mike age 12 city tianjin
OK
127.0.0.1:6379> hmget user:1 name city
1) "mike"
2) "tianjin"

（6） Judge field Whether there is
hexists key field
for example ,user：1 contain name Domain , So the return result is 1, Return when not included 0：
127.0.0.1:6379> hexists user:1 name
(integer) 1
（7） Get all field
hkeys key
hkeys The order should be hfields More appropriate , It returns all of the field, example
Such as ：
127.0.0.1:6379> hkeys user:1
1) "name"
2) "age"
3) "city"
（8） Get all value
hvals key
The following operation gets user：1 All value：
127.0.0.1:6379> hvals user:1
1) "mike"
2) "12"
3) "tianjin"
（9） Get all field-value 

hgetall key
The following operation gets user：1 be-all field-value：
127.0.0.1:6379> hgetall user:1
1) "name"
2) "mike"
3) "age"
4) "12"
5) "city"
6) "tianjin"
Development tips
In the use of hgetall when , If there are many hash elements , There will be blocking Redis The possibility of .
If the developer just needs to get part of it field, have access to hmget, If you have to get everything
field-value, have access to hscan command , This command iterates over the hash type ,hscan Will be in
2.7 Section introduction .
（10）hincrby hincrbyfloat
hincrby key field
hincrbyfloat key field
hincrby and hincrbyfloat, It's like incrby and incrbyfloat command , But their work
The domain is filed.
（11） Calculation value String length of （ need Redis3.2 above ）
hstrlen key field
for example hget user：1name Of value yes tom, that hstrlen The return result is 3：
127.0.0.1:6379> hstrlen user:1 name
(integer) 3

surface 2-3 Is the time complexity of hash type commands , Developers can refer to this table to select the appropriate
The order of .

 2.3.2　 Internal encoding
There are two types of internal encoding for hash types ：
·ziplist（ Compressed list ）： When the number of hash type elements is less than hash-max-ziplist-entries
To configure （ Default 512 individual ）、 At the same time, all values are less than hash-max-ziplist-value To configure （ Default 64
byte ） when ,Redis Will use ziplist As an internal implementation of hash ,ziplist Use more compact
Structure realizes the continuous storage of multiple elements , So in terms of saving memory hashtable Better .
·hashtable（ Hashtable ）： When the hash type cannot be satisfied ziplist The condition of ,Redis Can make
use hashtable As an internal implementation of hash , Because at this time ziplist The efficiency of reading and writing will decrease , and
hashtable The time complexity of reading and writing is O（1）.
The following example demonstrates the internal encoding of a hash type , And the corresponding changes .
1） When field The number is small and there is no big one value when , The internal code is ziplist：
127.0.0.1:6379> hmset hashkey f1 v1 f2 v2
OK
127.0.0.1:6379> object encoding hashkey
"ziplist"
2.1） When there is value Greater than 64 byte , The internal code will be written by ziplist Turn into hashtable：
127.0.0.1:6379> hset hashkey f3 "one string is bigger than 64 byte... Ignore ..."
OK
127.0.0.1:6379> object encoding hashkey
"hashtable"
2.2） When field The number exceeds 512, Internal coding will also be done by ziplist Turn into hashtable：
127.0.0.1:6379> hmset hashkey f1 v1 f2 v2 f3 v3 ... Ignore ... f513 v513
OK
127.0.0.1:6379> object encoding hashkey
"hashtable"
 

2.3.3　 Use scenarios
chart 2-15 Two pieces of user information recorded for a relational data table , The user's properties are the columns of the table ,
Each user information as a line .
If you store it in a hash type , Pictured 2-16 Shown .

Instead of caching user information using string serialization , Hash types become more intuitive , also
102
It will be more convenient to update . Each user's id Defined as the key suffix , How to field-
value Attribute corresponding to each user , Similar to the following pseudo code ：

UserInfo getUserInfo(long id){
//   user  id  As  key  suffix 
userRedisKey = "user:info:" + id;
//   Use  hgetall  Get all user information mapping relationships 
userInfoMap = redis.hgetAll(userRedisKey);
UserInfo userInfo;
if (userInfoMap != null) {
//   Convert the mapping relationship to  UserInfo
userInfo = transferMapToUserInfo(userInfoMap);
} else {
//   from  MySQL  Get user information 
userInfo = mysql.get(id);
//   take  userInfo  Change to mapping relation using  hmset  Save to  Redis  in 
redis.hmset(userRedisKey, transferUserInfoToMap(userInfo));
//   Add expiration time 
redis.expire(userRedisKey, 3600);
}
return userInfo;
}

But it should be noted that there are two differences between hash type and relational database ：
· Hash type is sparse , And relational databases are completely structured , For example, hash type
Each key can have a different field, Once a new column is added to a relational database , All lines should be
Its setting value （ Even for NULL）, Pictured 2-17 Shown .
· Relational databases can do complex relational queries , and Redis To simulate complex relational queries
Development is difficult , Maintenance costs are high . 

  Developers need to understand the characteristics of both , In order to use the right technology in the right scene
Technique . up to now , We've been able to cache user information in three ways , Three methods are given below
Implementation method and analysis of advantages and disadvantages of the case .
1） Native string type ： One key per attribute .

set user:1:name tom
set user:1:age 23
set user:1:city beijing

advantage ： Simple and intuitive , Each property supports update operations .
shortcoming ： Too many keys , Large memory consumption , At the same time, the cohesion of user information is relatively poor ,
So this kind of solution will not be used in the production environment .

2） Serialized string type ： Serialize user information and save it with a key .
set user:1 serialize(userInfo)
advantage ： Simplify programming , If the reasonable use of serialization can improve the efficiency of memory use .
shortcoming ： Serialization and deserialization have some overhead , At the same time, every time the attribute is updated, the full
Part of the data is taken out for deserialization , Update and serialize to Redis in .
3） Hash type ： Each user attribute uses a pair of field-value, But with only one key
save .
hmset user:1 name tomage 23 city beijing
advantage ： Simple and intuitive , If used reasonably, it can reduce the use of memory space .
shortcoming ： To control the hash in ziplist and hashtable Conversion of two internal codes ,hashtable Meeting
Consume more memory .