MySQL performance optimization index function, hidden, prefix, hash index usage (2)

Then the last issue said ,MYSQL 8 Is indexed in a relatively MYSQL 5.7 There has been a lot of progress , One of the most prominent is the addition of functional indexes , When did this function start .

8.013 , Yes MYSQL 8.013 Functional indexes have been added in version , Here's an example

create table function_test (id bigint primary key auto_increment,
                            name varchar(20) not null,
                            birth_day datetime);

mysql> insert into function_test (name,birrth_day) values ('Tim','2018-09-09');
Query OK, 1 row affected (0.01 sec)

mysql> insert into function_test (name,birrth_day) values ('Sam','2020-01-09');
Query OK, 1 row affected (0.01 sec)

mysql> insert into function_test (name,birrth_day) values ('Pam','2021-01-09');
Query OK, 1 row affected (0.00 sec)

Alter table function_test add index month(birrth_day);

We actually get an index with functions
CREATE TABLE `function_test` (
  `id` bigint NOT NULL AUTO_INCREMENT,
  `name` varchar(20) NOT NULL,
  `birrth_day` datetime DEFAULT NULL,
  PRIMARY KEY (`id`),
  KEY `month` (`birrth_day`)
) ENGINE=InnoDB AUTO_INCREMENT=4 DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci |

explain select birrth_day from function_test where month(birrth_day) = 5;

explain select birrth_day from function_test where month(birrth_day) = 5;

Based on this function ,MYSQL The version of should be in 8.013 Above version .

The prefix index is MYSQL Unique features , But there are scenarios when using , That is to say, the amount type to which the prefix index is first added is the character type , The data can be searched in detail through the value of the field in the front part of the field , Its main purpose is to reduce the index bytes at the same time , It also improves efficiency .

Let's give an example

create  table prefix_test (id bigint primary key auto_increment,name varchar(20) not null, type_comments varchar(2000));

mysql> insert into prefix_test (name,type_comments) values ('Tim','1_ I don't like ');
Query OK, 1 row affected (0.01 sec)

mysql> insert into prefix_test (name,type_comments) values ('Sam','3_ I am very fond of ');
Query OK, 1 row affected (0.01 sec)

mysql> insert into prefix_test (name,type_comments) values ('Semon','2_ have no feelings ');
Query OK, 1 row affected (0.00 sec)

mysql> alter table prefix_test add index (type_comments(1));
Query OK, 0 rows affected (0.07 sec)
Records: 0  Duplicates: 0  Warnings: 0

explain select name from prefix_test where type_comments like '1%';


It can also be seen from the above example , In fact, we just need to give 2000 Character , One of them 1 Characters to index , You can complete some quick searches .

Or it can be done in this way GROUP BY   And so on .

After all this is MYSQL Hidden index for , This is not a new feature , There are similar index types in other databases , In fact, the hidden index is used to verify the validity of the index for some queries , And let him go online or offline .

mysql>
mysql> alter table prefix_test add index (type_comments(1)) invisible;
Query OK, 0 rows affected (0.08 sec)
Records: 0  Duplicates: 0  Warnings: 0

mysql> show create table prefix_test;
                                                                                                                                                                                      
CREATE TABLE `prefix_test` (
  `id` bigint NOT NULL AUTO_INCREMENT,
  `name` varchar(20) NOT NULL,
  `type_comments` varchar(2000) DEFAULT NULL,
  PRIMARY KEY (`id`),
  KEY `type_comments` (`type_comments`(1)) *!80000 INVISIBLE */
) ENGINE=InnoDB AUTO_INCREMENT=4 DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci |



mysql> explain select name from prefix_test where type_comments like '1%';

| id | select_type | table       | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |

|  1 | SIMPLE      | prefix_test | NULL       | ALL  | NULL          | NULL | NULL    | NULL |    3 |    33.33 | Using where |

1 row in set, 1 warning (0.00 sec)

The figure above shows that although the new index is created, the invisible The state of is unusable . If you want to use it, you can immediately let it into visible The state of .

And finally hash Indexes , In the last article, I mentioned that I have established HASH The method solves the problem that multiple fields in the equivalent index establish a joint index , Let's take an example to see how to operate .

mysql> alter table hash_test add hash_col varchar(20) generated always as (crc32(concat(type1,type2)));
Query OK, 0 rows affected (0.03 sec)
Records: 0  Duplicates: 0  Warnings: 0

mysql> select * from hash_test;
+----+------+-------+-------+------------+
| id | name | type1 | type2 | hash_col   |
+----+------+-------+-------+------------+
|  1 | Tim  | 10    | 20    | 2249310613 |
+----+------+-------+-------+------------+
1 row in set (0.00 sec)

mysql> insert into hash_test (name,type1,type2) values ('Sam','20','30');
Query OK, 1 row affected (0.00 sec)

mysql> select * from hash_test;
+----+------+-------+-------+------------+
| id | name | type1 | type2 | hash_col   |
+----+------+-------+-------+------------+
|  1 | Tim  | 10    | 20    | 2249310613 |
|  2 | Sam  | 20    | 30    | 2378129210 |
+----+------+-------+-------+------------+
2 rows in set (0.00 sec)

Now we can give hash Index the virtual fields of

So the next time we check , You can use it directly hash_col As the index field for the lookup , Instead of caring about ytype1 type2 If there is 50 Fields to be indexed , Then you can ignore , Just one field can express N Values for fields , In this way, the query speed is fast , At the same time, your index has only one at any time , The problem of space efficiency has been solved .