mysql 简单的sql优化示例[不定时更新]

Posted 2020-09-08 duenboa

对于慢sql的分析步骤:

1) desc|explain sql 查看执行计划, 对于type很慢的, 分析是否建立了对应字段的索引

2) 进行排除法, 把子查询抽离出来, 单独执行,定位慢查询是哪个子查询导致的. 专门进行优化.

 

脚本:

CREATE TABLE stu (

id BIGINT(20) PRIMARY KEY AUTO_INCREMENT COMMENT ‘主键‘,

NAME VARCHAR(50) NOT NULL DEFAULT ‘‘ COMMENT ‘名称‘,

phone BIGINT(20) NULL COMMENT ‘电话‘,

UNIQUE KEY idx_phone (phone)

)ENGINE INNODB DEFAULT CHARSET utf8 COMMENT ‘测试_学生表‘;

 

CREATE TABLE course (

id BIGINT(20) PRIMARY KEY AUTO_INCREMENT COMMENT ‘主键‘,

NAME VARCHAR(50) NOT NULL DEFAULT ‘‘ COMMENT ‘名称‘

)ENGINE INNODB DEFAULT CHARSET utf8 COMMENT ‘测试_课程表‘;

 

CREATE TABLE stu_course (

id BIGINT(20) PRIMARY KEY AUTO_INCREMENT COMMENT ‘主键‘,

stu_id BIGINT(20) NOT NULL COMMENT ‘学生id‘,

cous_id BIGINT(20) NOT NULL COMMENT ‘课程id‘,

score INT(11) NOT NULL DEFAULT 0 COMMENT ‘分数‘,

KEY idx_stu_id (stu_id),

KEY idx_cous_id (cous_id)

)ENGINE INNODB DEFAULT CHARSET utf8 COMMENT ‘测试_学生课程关联表‘;

 

1. OR 变 UNION ALL

前:

SELECT id FROM stu WHERE id=10 OR id=20

后:

SELECT id FROM stu WHERE id=10

UNION ALL

SELECT id FROM stu WHERE id=20

 

2. IN (子查询) 变 INNER JOIN

前:

SELECT id FROM stu WHERE id IN (SELECT stu_id FROM stu_course)

后:

SELECT id FROM stu s

INNER JOIN (SELECT stu_id FROM stu_course) tmp ON s.id = tmp.stu_id

 

3. NOT IN (子查询) 变 LEFT JOIN

前:

SELECT id FROM stu WHERE id NOT IN (SELECT stu_id FROM stu_course)

后:

SELECT id FROM stu s

LEFT JOIN (SELECT stu_id FROM stu_course) tmp ON s.id = tmp.stu_id

WHERE tmp.stu_id IS NULL

 

4. IN 变 EXISTS

前:

SELECT id FROM stu WHERE id IN (SELECT stu_id FROM stu_course)

后:

SELECT id FROM stu s WHERE EXISTS (SELECT sc.stu_id FROM stu_course sc WHERE sc.stu_id = s.id)