在开始博客之前,还是同样的给一个大概的目录结构,实则即为一般MySQL的优化步骤

1、查看SQL的执行频率—————使用show status命令

2、定位哪些需要优化的SQL————通过慢查询记录+show processlist命令查看当前线程

3、分析为什么SQL执行效率低————使用explain/desc命令分析

  • 相关列简单解释:type、table、select_type…

4、对症下药采取优化措施———–举例采取index进行优化

  • 如何使用索引?
  • 使用索引应该注意的事项
  • 查看索引使用情况

主要参考资料:《深入浅出MySQL》,https://dev.mysql.com/doc/refman/8.0/en/statement-optimization.html

 


 

一、查看SQL执行频率

  使用show [session|gobal] status命令了解SQL执行频率、线程缓存内的线程的数量、当前打开的连接的数量、获得的表的锁的次数等。

比如执行show status like \’Com_%\’查看每个语句执行的次数即频率,其中Com_xxx中xxx表示就是语句,比如Com_select:执行select操作的次数。

 1 mysql> use test;
 2 Database changed
 3 mysql> show status like \'Com_%\';
 4 +-----------------------------+-------+
 5 | Variable_name               | Value |
 6 +-----------------------------+-------+
 7 | Com_admin_commands          | 0     |
 8 | Com_assign_to_keycache      | 0     |
 9 | Com_alter_db                | 0     |
10 | Com_alter_db_upgrade        | 0     |
11 | Com_alter_event             | 0     |
12 | Com_alter_function          | 0     |
13 | Com_alter_instance          | 0     |
14 | Com_alter_procedure         | 0     |
15 | Com_alter_server            | 0     |
16 | Com_alter_table             | 0     |
17 | Com_alter_tablespace        | 0     |
18 | Com_alter_user              | 0     |
19 | Com_analyze                 | 0     |
20 | Com_begin                   | 0     |
21 | Com_binlog                  | 0     |
22 | Com_call_procedure          | 0     |
23 | Com_change_db               | 2     |
24 | Com_change_master           | 0     |
25 | Com_change_repl_filter      | 0     |
26 | Com_check                   | 0     |
27 | Com_checksum                | 0     |
28 | Com_commit                  | 0     |
29 | Com_create_db               | 0     |
30 | Com_create_event            | 0     |
31 | Com_create_function         | 0     |
32 | Com_create_index            | 0     |
  ..............................

比如执行show status like \’slow_queries\’查看慢查询次数(黑人问号??什么是慢查询呢?就是通过设置查询时间阈值long_query_time(0-10s)并打开开关show_query_log(1=OFF/0=ON),当超过这个阈值的查询都称之为慢查询,通常用来划分执行SQL效率)

mysql> show status like \'slow_queries\';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| Slow_queries  | 0     |
+---------------+-------+
1 row in set

比如执行show status like \’uptime\’查看服务工作时间(即运行时间)

mysql> show status like \'uptime\';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| Uptime        | 21645 |
+---------------+-------+
1 row in set

比如执行show status like \’connections\’查看MySQL连接数:

mysql> show status like \'connections\';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| Connections   | 6     |
+---------------+-------+
1 row in set

  通过show [session|gobal] status命令很清楚地看到哪些SQL执行效率不如人意,但是具体是怎么个不如意法,还得继续往下看,使用EXPLAIN命令分析具体的SQL语句

 二、定位效率低的SQL

  上面也提到过慢查询这个概念主要是用来划分效率低的SQL,但是慢查询是在整个查询结束后才记录的,所以光是靠慢查询日志是跟踪不了效率低的SQL。一般有两种方式定位效率低的SQL:

  1、通过慢查询日志查看效率低的SQL语句,慢查询日志是通过show_query_log_file指定存储路径的,里面记录所有超过long_query_time的SQL语句(关于日志的查看,日后再一步研究学习),但是需要慢查询日志的产生是在查询结束后才有的。

  2、通过show processlist命令查看当前MySQL进行的线程,可以看到线程的状态信息

mysql> show processlist;
+----+------+-----------------+------+---------+------+----------+------------------+
| Id | User | Host            | db   | Command | Time | State    | Info             |
+----+------+-----------------+------+---------+------+----------+------------------+
|  2 | root | localhost:58377 | NULL | Sleep   | 2091 |          | NULL             |
|  3 | root | localhost:58382 | test | Sleep   | 2083 |          | NULL             |
|  4 | root | localhost:58386 | test | Sleep   | 2082 |          | NULL             |
|  5 | root | localhost:59092 | test | Query   |    0 | starting | show processlist |
+----+------+-----------------+------+---------+------+----------+------------------+
4 rows in set

  其中主要的是state字段,表示当前SQL语句线程的状态,如Sleeping 表示正在等待客户端发送新请求,Sending data把查询到的data结果发送给客户端等等,具体请看https://dev.mysql.com/doc/refman/8.0/en/general-thread-states.html

三、 查看分析效率低的SQL

  MYSQL 5.6.3以前只能EXPLAIN SELECT; MYSQL5.6.3以后就可以EXPLAIN SELECT,UPDATE,DELETE,现在我们先创建一个user_table的表,之后分析select* from user where name=\’\’语句

mysql> create table user(id int, name varchar(10),password varchar(32),primary key(id))engine=InnoDB;
Query OK, 0 rows affected

之后插入三条数据:

mysql> insert into user values(1,\'Zhangsan\',replace(UUID(),\'-\',\'\')),(2,\'Lisi\',replace(UUID(),\'-\',\'\')),(3,\'Wangwu\',replace(UUID(),\'-\',\'\'));
Query OK, 3 rows affected
Records: 3  Duplicates: 0  Warnings: 0
mysql> select* from user;
+----+----------+----------------------------------+
| id | name     | password                         |
+----+----------+----------------------------------+
|  1 | Zhangsan | 2d7284808e5111e8af74201a060059ce |
|  2 | Lisi     | 2d73641c8e5111e8af74201a060059ce |
|  3 | Wangwu   | 2d73670c8e5111e8af74201a060059ce |
+----+----------+----------------------------------+
3 rows in set

下面以分析select*from user where name=\’Lisi\’语句为例:

mysql> explain select*from user where name=\'Lisi\';
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | user  | NULL       | ALL  | NULL          | NULL | NULL    | NULL |    3 |    33.33 | Using where |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set

 

下面讲解select_type等常见列的含义的:

(1)select_type:表示SELECT的类型,主要有:

  • SIMPLE:简单表,没有表连接或者子查询
  • PRIMARY:主查询,即最外城的查询
  • UNION:UNION中的第二个或者后面的语句
  • SUBQUERY:子查询中的第一个SELECT

(2)table:结果输出的表

(3)type:表示表的连接类型,性能由好到差为:

  • system:常量表
  • const:单表中最多有一行匹配,比如primary key,unique index
  • eq_ref:多表连接中使用primary key,unique index
  • ref:使用普通索引
  • ref_or_null:与ref类似,但是包含了NULL查询
  • index_merge:索引合并优化
  • unique_subquery:in后面是一个查询主键字段的子查询
  • index_subquery:in后面是非唯一索引字段的子查询
  • range:单表中范围查看,使用like模糊查询
  • index:对于后面每一行都通过查询索引得到数据
  • all:表示全表查询

(3)possible_key:查询时可能使用的索引

(4)key:表示实际使用的索引

(5)key_len:索引字段的长度

(6)rows:查询时实际扫描的行数

(7)Extra:执行情况的说明和描述

(8)partitions:分区数目

(9)filtered:查询过滤的表占的百分比,比如这里查询的记录是name=Lisi的记录,占三条记录的33.3%

四、 关于索引的优化

1、使用索引优化的举例

  上个例子我们看到到执行explain select*from user where name=\’Lisi\’,扫描了3行(全部行数)使用了全表搜索all。如果实际业务中name是经常用到查询的字段(是指经常跟在where后的字段,不是select后的字段)并且数据量很大的情况呢?这时候就需要索引了(索引经常用到where后面的字段比select后面的字段效果更好,或者说就是要使用在where后面的字段上)

增加name前缀索引(这里只是举例,并没有选择最合适的前缀):

mysql> create index index_name on user(name(2));
Query OK, 0 rows affected
Records: 0  Duplicates: 0  Warnings: 0

执行explain分析

mysql> explain select*from user where name = \'Lisi\';
+----+-------------+-------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key        | key_len | ref   | rows | filtered | Extra       |
+----+-------------+-------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
|  1 | SIMPLE      | user  | NULL       | ref  | index_name    | index_name | 9       | const |    1 |      100 | Using where |
+----+-------------+-------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
1 row in set

  可以看到type变为ref、rows降为1(实际上只要使用了索引都是1),filtered过滤百分比为100%,实际用到的索引为index_name。如果数据量很大的话使用索引就是很好的优化措施,对于如何选择索引,什么时候用索引,我做出了如下总结:

2、如何高效使用索引?

  (1) 创建多列索引时,只要查询条件中用到最左边的列,索引一般都会被用到

  我们创建一张没有索引的表user_1:

mysql> show create table 
user_1;
+--------+--------------------------------------------------------------------------------------------------------------------------+
| Table  | Create Table                                                                                                             |
+--------+--------------------------------------------------------------------------------------------------------------------------+
| user_1 | CREATE TABLE `user_1` (
  `id` int(11) DEFAULT NULL,
  `name` varchar(10) DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8 |
+--------+--------------------------------------------------------------------------------------------------------------------------+
 1 row in set

 之后同样插入数据:

mysql> select *from user_1;
+----+----------+
| id | name     |
+----+----------+
|  1 | Zhangsan |
|  2 | Lisi     |
+----+----------+
2 rows in set

 创建多列索引index_id_name

mysql> create index index_id_name on user_1(id,name);
Query OK, 0 rows affected
Records: 0  Duplicates: 0  Warnings: 0

 实验查询explain分析name与id

mysql> explain select * from user_1 where id=1;
+----+-------------+--------+------------+------+---------------+---------------+---------+-------+------+----------+-------------+
| id | select_type | table  | partitions | type | possible_keys | key           | key_len | ref   | rows | filtered | Extra       |
+----+-------------+--------+------------+------+---------------+---------------+---------+-------+------+----------+-------------+
|  1 | SIMPLE      | user_1 | NULL       | ref  | index_id_name | index_id_name | 5       | const |    1 |      100 | Using index |
+----+-------------+--------+------------+------+---------------+---------------+---------+-------+------+----------+-------------+
1 row in set

mysql> explain select * from user_1 where name=\'Lisi\';
+----+-------------+--------+------------+-------+---------------+---------------+---------+------+------+----------+--------------------------+
| id | select_type | table  | partitions | type  | possible_keys | key           | key_len | ref  | rows | filtered | Extra                    |
+----+-------------+--------+------------+-------+---------------+---------------+---------+------+------+----------+--------------------------+
|  1 | SIMPLE      | user_1 | NULL       | index | NULL          | index_id_name | 38      | NULL |    2 |       50 | Using where; Using index |
+----+-------------+--------+------------+-------+---------------+---------------+---------+------+------+----------+--------------------------+
1 row in set

  可以看到使用最左列id的时候,rows为1,并且Extra明确使用了index,key的值为id_name_index,type的值为ref,而where不用到id,而是name的话,rows的值为2。filtered为50%,虽然key是index_id_name,但是表明是索引(个人理解,应该不太准确)

  (2) 使用like的查询,只有%不是第一个字符并且%后面是常量的情况下,索引才可能会被使用。

   执行explain select *from user where name like ‘%Li’后type为ALLkey的值为NULL,执行explain select *from user where name like ‘Li%’后key值不为空为index_name。

mysql> explain select*from user where name like \'%Li\';
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | user  | NULL       | ALL  | NULL          | NULL | NULL    | NULL |    3 |    33.33 | Using where |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set
mysql> explain select*from user where name like \'Li%\';
+----+-------------+-------+------------+-------+---------------+------------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type  | possible_keys | key        | key_len | ref  | rows | filtered | Extra       |
+----+-------------+-------+------------+-------+---------------+------------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | user  | NULL       | range | index_name    | index_name | 9       | NULL |    1 |      100 | Using where |
+----+-------------+-------+------------+-------+---------------+------------+---------+------+------+----------+-------------+
1 row in set

  (3) 如果对打的文本进行搜索,使用全文索引而不是用like ‘%…%’(只有MyISAM支持全文索引)

  (4) 如果列名是索引,使用column_name is null将使用索引

mysql> explain select*from user where name is null;
+----+-------------+-------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key        | key_len | ref   | rows | filtered | Extra       |
+----+-------------+-------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
|  1 | SIMPLE      | user  | NULL       | ref  | index_name    | index_name | 9       | const |    1 |      100 | Using where |
+----+-------------+-------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
1 row in set

mysql> explain select*from user where password
 is null;
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | user  | NULL       | ALL  | NULL          | NULL | NULL    | NULL |    3 |    33.33 | Using where |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
1 row in set

3、哪些情况下即使有索引也用不到?

  (1) MySQL使用MEMORY/HEAP引擎(使用的HASH索引),并且WHERE条件中不会使用”=”,in等进行索引列,那么不会用到索引(这是关于引擎部分特点,之后会介绍)。

  (2) 用OR分隔开的条件,如果OR前面的条件中的列有索引,而后面的列没有索引,那么涉及到的列索引不会被使用。

  执行命令show index from user可以看出password字段并没有使用任何索引,而id使用了两个索引,但是where id=1 or password=\’2d7284808e5111e8af74201a060059ce\’ 导致没有使用id列的primary索引与id_name_index索引

mysql> show index from user;
+-------+------------+---------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| Table | Non_unique | Key_name      | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | Index_comment |
+-------+------------+---------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| user  |          0 | PRIMARY       |            1 | id          | A         |           3 | NULL     | NULL   |      | BTREE      |         |               |
| user  |          1 | index_name    |            1 | name        | A         |           3 |        2 | NULL   | YES  | BTREE      |         |               |
| user  |          1 | id_name_index |            1 | id          | A         |           3 | NULL     | NULL   |      | BTREE      |         |               |
| user  |          1 | id_name_index |            2 | name        | A         |           3 | NULL     | NULL   | YES  | BTREE      |         |               |
+-------+------------+---------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
4 rows in set

mysql> explain select*from user where id=1 or password=\'2d7284808e5111e8af74201a060059ce\';
+----+-------------+-------+------------+------+-----------------------+------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys         | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+-------+------------+------+-----------------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | user  | NULL       | ALL  | PRIMARY,id_name_index | NULL | NULL    | NULL |    3 |    55.56 | Using where |
+----+-------------+-------+------------+------+-----------------------+------+---------+------+------+----------+-------------+
1 row in set

  (3) 不是用到复合索引中的第一列即最左边的列的话,索引就不起作用(上面已经介绍)。

  (4) 如果like是以%开头的(上面已经介绍)

  (5) 如果列类型是字符串,那么where条件中字符常量值不用’’引号引起来的话,那就不会失去索引效果,这是因为MySQL会把输入的常量值进行转换再使用索引。

  select * from user_1 where name =250,其中name的索引为name_index,并且是varchar字符串类型,但是并没有将250用引号变成’250’,那么explain之后的ref仍然为NULL,rows为3

mysql> show index from user_1;
+--------+------------+---------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| Table  | Non_unique | Key_name      | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | Index_comment |
+--------+------------+---------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| user_1 |          1 | index_id_name |            1 | id          | A         |           2 | NULL     | NULL   | YES  | BTREE      |         |               |
| user_1 |          1 | index_id_name |            2 | name        | A         |           2 | NULL     | NULL   | YES  | BTREE      |         |               |
| user_1 |          1 | name_index    |            1 | name        | A         |           3 |        5 | NULL   | YES  | BTREE      |         |               |
+--------+------------+---------------+--------------+-------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
3 rows in set

mysql> explain select*from user_1 where name=250;
+----+-------------+--------+------------+-------+---------------+---------------+---------+------+------+----------+--------------------------+
| id | select_type | table  | partitions | type  | possible_keys | key           | key_len | ref  | rows | filtered | Extra                    |
+----+-------------+--------+------------+-------+---------------+---------------+---------+------+------+----------+--------------------------+
|  1 | SIMPLE      | user_1 | NULL       | index | name_index    | index_id_name | 38      | NULL |    3 |    33.33 | Using where; Using index |
+----+-------------+--------+------------+-------+---------------+---------------+---------+------+------+----------+--------------------------+
1 row in set

mysql> explain select*from user_1 where name=\'250\';
+----+-------------+--------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
| id | select_type | table  | partitions | type | possible_keys | key        | key_len | ref   | rows | filtered | Extra       |
+----+-------------+--------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
|  1 | SIMPLE      | user_1 | NULL       | ref  | name_index    | name_index | 18      | const |    1 |      100 | Using where |
+----+-------------+--------+------------+------+---------------+------------+---------+-------+------+----------+-------------+
1 row in set

 

4、查看索引的使用情况

执行show status like Handler_read%’可以看到一个值Handler_read_key,它代表一行被索引值读的次数,如果值很低说明增加索引得到的性能改善不高,因为索引并不经常使用。

mysql> show status like \'Handler_read%\' ;
+-----------------------+-------+
| Variable_name         | Value |
+-----------------------+-------+
| Handler_read_first    | 3     |
| Handler_read_key      | 5     |
| Handler_read_last     | 0     |
| Handler_read_next     | 0     |
| Handler_read_prev     | 0     |
| Handler_read_rnd      | 0     |
| Handler_read_rnd_next | 20    |
+-----------------------+-------+
7 rows in set

(1)Handler_read_first:索引中第一条被读的次数。如果较高,它表示服务器正执行大量全索引扫描;

(2)Handler_read_key:如果索引正在工作,这个值代表一个行被索引值读的次数,如果值越低,表示索引得到的性能改善不高,因为索引不经常使用。

(3)Handler_read_next :按照键顺序读下一行的请求数。如果你用范围约束或如果执行索引扫描来查询索引列,该值增加。

(4)Handler_read_prev:按照键顺序读前一行的请求数。该读方法主要用于优化ORDER BY … DESC。

(5)Handler_read_rnd :根据固定位置读一行的请求数。如果你正执行大量查询并需要对结果进行排序该值较高。你可能使用了大量需要MySQL扫描整个表的查询或你的连接没有正确使用键。这个值较高,意味着运行效率低,应该建立索引来补救。

(6)Handler_read_rnd_next:在数据文件中读下一行的请求数。如果你正进行大量的表扫描,该值较高。通常说明你的表索引不正确或写入的查询没有利用索引。

   注:以上6点来自于网络总结,其中比较重要的两个参数是Handler_read_key与Handler_read_rnd_next。

版权声明:本文为jian0110原创文章,遵循 CC 4.0 BY-SA 版权协议,转载请附上原文出处链接和本声明。
本文链接:https://www.cnblogs.com/jian0110/p/9356347.html