Use and optimization of MySQL composite index (multi column index)

Catalog

1、 Multi column index

2、 Test cases and process

2.1 Create a test database and data table

2.2 Add two single column indexes

2.3 Query a piece of data using the index of two columns

2.4 View execution plan

2.5 Then delete the above index , Add multi column index

2.6 Query again

3、 Use order of multi column index

3.1 How to choose to build a composite index , Column order

3.2 Rules for using composite indexes

---

1、 Multi column index

We often hear some people say " hold WHERE The columns in the condition are indexed ", In fact, this suggestion is very wrong .

Building a separate index on multiple columns in most cases does not improve MySQL Query performance of .MySQL stay 5.0 After that, a new technology called “ Index merging ”（index merge） The strategy of , To some extent, multiple single column indexes on the table can be used to locate the specified row . But when the server performs joint operations on multiple indexes , It usually takes a lot of CPU And memory resources in the algorithm cache 、 Sorting and merging operations , Especially when the selectivity of some indexes is not high , When you need to merge and scan a large amount of data . This is the time , We need a multi column index .

2、 Test cases and process

2.1 Create a test database and data table

CREATE DATABASE IF NOT EXISTS db_test default charset utf8 COLLATE utf8_general_ci; 
use db_test;
 
CREATE TABLE payment (  
    id        INT UNSIGNED NOT NULL AUTO_INCREMENT,  
    staff_id  INT UNSIGNED NOT NULL,
    customer_id INT UNSIGNED NOT NULL, 
    PRIMARY KEY (id)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;

Use stored procedure to insert 1000w Row random data ( The table engine can be set to MyISAM, Then change it to InnoDB)

DROP PROCEDURE IF EXISTS add_payment;  
DELIMITER //
    create PROCEDURE add_payment(in num INT)
    BEGIN
        DECLARE rowid INT DEFAULT 0;
        SET @exesql = 'INSERT INTO payment(staff_id, customer_id) values (?, ?)';
        WHILE rowid < num DO
            SET @staff_id = (1 + FLOOR(5000*RAND()) ); 
            SET @customer_id = (1 + FLOOR(500000*RAND()));
            SET rowid = rowid + 1;
            prepare stmt FROM @exesql;
            EXECUTE stmt USING @staff_id, @customer_id;            
        END WHILE;
    END //
DELIMITER ;

2.2 Add two single column indexes

( The execution process takes some time , It is suggested to implement sentence by sentence ):

ALTER TABLE `payment` ADD INDEX idx_customer_id(`customer_id`);
ALTER TABLE `payment` ADD INDEX idx_staff_id(`staff_id`);

2.3 Query a piece of data using the index of two columns

select count(*) from payment where staff_id = 2205 AND customer_id = 93112;

2.4 View execution plan

mysql> explain select count(*)  from payment  where staff_id =  2205  AND customer_id =  93112;
+----+-------------+---------+-------------+------------------------------+------------------------------+---------+------+-------+-------------------------------------------------------------------------+
| id | select_type | table   | type        | possible_keys                | key                          | key_len | ref  | rows  | Extra                                                                   |
+----+-------------+---------+-------------+------------------------------+------------------------------+---------+------+-------+-------------------------------------------------------------------------+
|  1 | SIMPLE      | payment | index_merge | idx_customer_id,idx_staff_id | idx_staff_id,idx_customer_id | 4,4     | NULL | 11711 | Using intersect(idx_staff_id,idx_customer_id); Using where; Using index |
+----+-------------+---------+-------------+------------------------------+------------------------------+---------+------+-------+-------------------------------------------------------------------------+
1 row in set (0.00 sec)

You can see type yes index_merge,Extra Middle tip Using intersect(idx_staff_id,idx_customer_id); This is index merging , Use two indexes , Then merge the two results ( Take intersection or union or both )

Query results ：

mysql> select count(*)  from payment  where staff_id =  2205  AND customer_id =  93112 ;
+----------+
| count(*) |
+----------+
|   178770 |
+----------+
1 row in set (0.12 sec)

2.5 Then delete the above index , Add multi column index

ALTER TABLE payment DROP INDEX idx_customer_id;
ALTER TABLE payment DROP INDEX idx_staff_id;
ALTER TABLE `payment` ADD INDEX idx_customer_id_staff_id(`customer_id`, `staff_id`);

Be careful , Multi column index pays much attention to the order of index columns （ because customer_id More selective , So put it in front ）.

2.6 Query again

mysql> select count(*)  from payment  where staff_id =  2205  AND customer_id =  93112;
+----------+
| count(*) |
+----------+
|   178770 |
+----------+
1 row in set (0.05 sec)

Find queries that are faster with multi column indexes （ The amount of data here is still small , It is more obvious when you are older ）.

3、 Use order of multi column index

3.1 How to choose to build a composite index , Column order

The order of columns in a multi column index is critical , There is a rule of thumb about how to choose the column order of an index ： Put the most selective column at the forefront of the index （ But not absolutely ）. The rule of thumb considers global cardinality and selectivity , Instead of a specific query ：

mysql> select count(DISTINCT staff_id) / count(*) AS staff_id_selectivity, count(DISTINCT customer_id) / count(*) AS customer_id_selectivity, count(*) from payment\G;
*************************** 1. row ***************************
   staff_id_selectivity: 0.0005
customer_id_selectivity: 0.0500
               count(*): 10000000
1 row in set (6.29 sec)

customer_id It's more selective , So take it as the first place in the index column .

3.2 Rules for using composite indexes

Indexes can be understood as ordered data structures . Composite index can be understood as , such as （a,b,c）,abc It's all in order , In any paragraph a Below b It's all in order , Any paragraph b below   c It's all in order ;

The effective rule is ： Use from front to back in turn , If an index in the middle is not used , So the index section in front of the breakpoint works , The index after the breakpoint doesn't work ; 

such as ：

where a=3 and b=45 and c=5 ....  This three index sequence uses no breakpoints in the middle , It all works ;
where a=3 and c=5...  In this case b It's a breakpoint ,a It worked ,c There is no effect 
where b=3 and c=4...  In this case a It's a breakpoint , stay a None of the later indexes worked , This method of writing union index has no effect ;
where b=45 and a=3 and c=5 ....  This is the same as the first one , It all works ,abc Just use it , It's not about the order of writing 

（a,b,c） Examples of the use of multi column indexes , explain ：（a,b,c） Combine index and (a,c,b） It's different

(0)    select * from mytable where a=3 and b=5 and c=4;
abc All three indexes are in where It's used in the conditions , And it all worked 
(1)    select * from mytable where  c=4 and b=6 and a=3;
 This statement is listed just to show  mysql It's not that stupid ,where The order of the conditions inside will be changed before the query mysql Automatic optimization , The effect is the same as the sentence 
(2)    select * from mytable where a=3 and c=7;
a Use index ,b of no avail , therefore c There is no index effect 
(3)    select * from mytable where a=3 and b>7 and c=3;
a Yes ,b Also used. ,c Not used , This place b It's the range value , It's also a breakpoint , It just uses the index itself 
(4)    select * from mytable where b=3 and c=4;
 because a The index is not used , So here  bc No indexing effect 
(5)    select * from mytable where a>4 and b=7 and c=9;
a Yes   b Not used ,c Not used 
(6)    select * from mytable where a=3 order by b;
a Index is used ,b Index effect is also used in result sorting , I said before. ,a In any of the following paragraphs b It's in order 
(7)    select * from mytable where a=3 order by c;
a Index is used , But this place c No ranking effect , Because there's a break in the middle , Use  explain  You can see  filesort
(8)    select * from mytable where b=3 order by a;
b No index is used , Sorting a It doesn't have an indexing effect 

Original author ： Qin Tian

Original address ：https://blog.csdn.net/demored/article/details/123068511

（ Thanks to the original author for sharing ）