mysql四-2:多表查询
一 介绍
本节主题
- 多表连接查询
- 复合条件连接查询
- 子查询
准备表
#建表 create table department( id int, name varchar(20) ); create table employee( id int primary key auto_increment, name varchar(20), sex enum('male','female') not null default 'male', age int, dep_id int ); #插入数据 insert into department values (200,'技术'), (201,'人力资源'), (202,'销售'), (203,'运营'); insert into employee(name,sex,age,dep_id) values ('egon','male',18,200), ('alex','female',48,201), ('wupeiqi','male',38,201), ('yuanhao','female',28,202), ('liwenzhou','male',18,200), ('jingliyang','female',18,204) ; #查看表结构和数据 mysql> desc department; +-------+-------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +-------+-------------+------+-----+---------+-------+ | id | int(11) | YES | | NULL | | | name | varchar(20) | YES | | NULL | | +-------+-------------+------+-----+---------+-------+ mysql> desc employee; +--------+-----------------------+------+-----+---------+----------------+ | Field | Type | Null | Key | Default | Extra | +--------+-----------------------+------+-----+---------+----------------+ | id | int(11) | NO | PRI | NULL | auto_increment | | name | varchar(20) | YES | | NULL | | | sex | enum('male','female') | NO | | male | | | age | int(11) | YES | | NULL | | | dep_id | int(11) | YES | | NULL | | +--------+-----------------------+------+-----+---------+----------------+ mysql> select * from department; +------+--------------+ | id | name | +------+--------------+ | 200 | 技术 | | 201 | 人力资源 | | 202 | 销售 | | 203 | 运营 | +------+--------------+ mysql> select * from employee; +----+------------+--------+------+--------+ | id | name | sex | age | dep_id | +----+------------+--------+------+--------+ | 1 | egon | male | 18 | 200 | | 2 | alex | female | 48 | 201 | | 3 | wupeiqi | male | 38 | 201 | | 4 | yuanhao | female | 28 | 202 | | 5 | liwenzhou | male | 18 | 200 | | 6 | jingliyang | female | 18 | 204 | +----+------------+--------+------+--------+ 表department与employee
二 多表连接查询
#重点：外链接语法
SELECT 字段列表
FROM 表1 INNER|LEFT|RIGHT JOIN 表2
ON 表1.字段 = 表2.字段;
1 交叉连接：不适用任何匹配条件。生成笛卡尔积
mysql> select * from employee,department; +----+------------+--------+------+--------+------+--------------+ | id | name | sex | age | dep_id | id | name | +----+------------+--------+------+--------+------+--------------+ | 1 | egon | male | 18 | 200 | 200 | 技术 | | 1 | egon | male | 18 | 200 | 201 | 人力资源 | | 1 | egon | male | 18 | 200 | 202 | 销售 | | 1 | egon | male | 18 | 200 | 203 | 运营 | | 2 | alex | female | 48 | 201 | 200 | 技术 | | 2 | alex | female | 48 | 201 | 201 | 人力资源 | | 2 | alex | female | 48 | 201 | 202 | 销售 | | 2 | alex | female | 48 | 201 | 203 | 运营 | | 3 | wupeiqi | male | 38 | 201 | 200 | 技术 | | 3 | wupeiqi | male | 38 | 201 | 201 | 人力资源 | | 3 | wupeiqi | male | 38 | 201 | 202 | 销售 | | 3 | wupeiqi | male | 38 | 201 | 203 | 运营 | | 4 | yuanhao | female | 28 | 202 | 200 | 技术 | | 4 | yuanhao | female | 28 | 202 | 201 | 人力资源 | | 4 | yuanhao | female | 28 | 202 | 202 | 销售 | | 4 | yuanhao | female | 28 | 202 | 203 | 运营 | | 5 | liwenzhou | male | 18 | 200 | 200 | 技术 | | 5 | liwenzhou | male | 18 | 200 | 201 | 人力资源 | | 5 | liwenzhou | male | 18 | 200 | 202 | 销售 | | 5 | liwenzhou | male | 18 | 200 | 203 | 运营 | | 6 | jingliyang | female | 18 | 204 | 200 | 技术 | | 6 | jingliyang | female | 18 | 204 | 201 | 人力资源 | | 6 | jingliyang | female | 18 | 204 | 202 | 销售 | | 6 | jingliyang | female | 18 | 204 | 203 | 运营 | +----+------------+--------+------+--------+------+--------------+
2 内连接：只连接匹配的行
#找两张表共有的部分，相当于利用条件从笛卡尔积结果中筛选出了正确的结果 #department没有204这个部门，因而employee表中关于204这条员工信息没有匹配出来 mysql> select employee.id,employee.name,employee.age,employee.sex,department.name from employee inner join department on employee.dep_id=department.id; +----+-----------+------+--------+--------------+ | id | name | age | sex | name | +----+-----------+------+--------+--------------+ | 1 | egon | 18 | male | 技术 | | 2 | alex | 48 | female | 人力资源 | | 3 | wupeiqi | 38 | male | 人力资源 | | 4 | yuanhao | 28 | female | 销售 | | 5 | liwenzhou | 18 | male | 技术 | +----+-----------+------+--------+--------------+ #上述sql等同于 mysql> select employee.id,employee.name,employee.age,employee.sex,department.name from employee,department where employee.dep_id=department.id;
3 外链接之左连接：优先显示左表全部记录
#以左表为准，即找出所有员工信息，当然包括没有部门的员工 #本质就是：在内连接的基础上增加左边有右边没有的结果 mysql> select employee.id,employee.name,department.name as depart_name from employee left join department on employee.dep_id=department.id; +----+------------+--------------+ | id | name | depart_name | +----+------------+--------------+ | 1 | egon | 技术 | | 5 | liwenzhou | 技术 | | 2 | alex | 人力资源 | | 3 | wupeiqi | 人力资源 | | 4 | yuanhao | 销售 | | 6 | jingliyang | NULL | +----+------------+--------------+
4 外链接之右连接：优先显示右表全部记录
#以右表为准，即找出所有部门信息，包括没有员工的部门 #本质就是：在内连接的基础上增加右边有左边没有的结果 mysql> select employee.id,employee.name,department.name as depart_name from employee right join department on employee.dep_id=department.id; +------+-----------+--------------+ | id | name | depart_name | +------+-----------+--------------+ | 1 | egon | 技术 | | 2 | alex | 人力资源 | | 3 | wupeiqi | 人力资源 | | 4 | yuanhao | 销售 | | 5 | liwenzhou | 技术 | | NULL | NULL | 运营 | +------+-----------+--------------+
5 全外连接：显示左右两个表全部记录
全外连接：在内连接的基础上增加左边有右边没有的和右边有左边没有的结果 #注意：mysql不支持全外连接 full JOIN #强调：mysql可以使用此种方式间接实现全外连接 select * from employee left join department on employee.dep_id = department.id union select * from employee right join department on employee.dep_id = department.id ; #查看结果 +------+------------+--------+------+--------+------+--------------+ | id | name | sex | age | dep_id | id | name | +------+------------+--------+------+--------+------+--------------+ | 1 | egon | male | 18 | 200 | 200 | 技术 | | 5 | liwenzhou | male | 18 | 200 | 200 | 技术 | | 2 | alex | female | 48 | 201 | 201 | 人力资源 | | 3 | wupeiqi | male | 38 | 201 | 201 | 人力资源 | | 4 | yuanhao | female | 28 | 202 | 202 | 销售 | | 6 | jingliyang | female | 18 | 204 | NULL | NULL | | NULL | NULL | NULL | NULL | NULL | 203 | 运营 | +------+------------+--------+------+--------+------+--------------+ #注意 union与union all的区别：union会去掉相同的纪录
三 符合条件连接查询
#示例1：以内连接的方式查询employee和department表，并且employee表中的age字段值必须大于25,即找出年龄大于25岁的员工以及员工所在的部门
select employee.name,department.name from employee inner join department
on employee.dep_id = department.id
where age > 25;
#示例2：以内连接的方式查询employee和department表，并且以age字段的升序方式显示
select employee.id,employee.name,employee.age,department.name from employee,department
where employee.dep_id = department.id
and age > 25
order by age asc;
四 子查询
#1：子查询是将一个查询语句嵌套在另一个查询语句中。 #2：内层查询语句的查询结果，可以为外层查询语句提供查询条件。 #3：子查询中可以包含：IN、NOT IN、ANY、ALL、EXISTS 和 NOT EXISTS等关键字 #4：还可以包含比较运算符：= 、 !=、> 、<等
1 带IN关键字的子查询
#查询平均年龄在25岁以上的部门名
select id,name from department
where id in
(select dep_id from employee group by dep_id having avg(age) > 25);
#查看技术部员工姓名
select name from employee
where dep_id in
(select id from department where name='技术');
#查看不足1人的部门名
select name from department
where id in
(select dep_id from employee group by dep_id having count(id) <=1);
2 带比较运算符的子查询
#比较运算符：=、!=、>、>=、<、<=、<> #查询大于所有人平均年龄的员工名与年龄 mysql> select name,age from emp where age > (select avg(age) from emp); +---------+------+ | name | age | +---------+------+ | alex | 48 | | wupeiqi | 38 | +---------+------+ 2 rows in set (0.00 sec) #查询大于部门内平均年龄的员工名、年龄 select t1.name,t1.age from emp t1 inner join (select dep_id,avg(age) avg_age from emp group by dep_id) t2 on t1.dep_id = t2.dep_id where t1.age > t2.avg_age;
3 带EXISTS关键字的子查询
EXISTS关字键字表示存在。在使用EXISTS关键字时，内层查询语句不返回查询的记录。
而是返回一个真假值。True或False
当返回True时，外层查询语句将进行查询；当返回值为False时，外层查询语句不进行查询
#department表中存在dept_id=203，Ture
mysql> select * from employee
-> where exists
-> (select id from department where id=200);
+----+------------+--------+------+--------+
| id | name | sex | age | dep_id |
+----+------------+--------+------+--------+
| 1 | egon | male | 18 | 200 |
| 2 | alex | female | 48 | 201 |
| 3 | wupeiqi | male | 38 | 201 |
| 4 | yuanhao | female | 28 | 202 |
| 5 | liwenzhou | male | 18 | 200 |
| 6 | jingliyang | female | 18 | 204 |
+----+------------+--------+------+--------+
#department表中存在dept_id=205，False
mysql> select * from employee
-> where exists
-> (select id from department where id=204);
Empty set (0.00 sec)
练习：查询每个部门最新入职的那位员工
company.employee 员工id id int 姓名 emp_name varchar 性别 sex enum 年龄 age int 入职日期 hire_date date 岗位 post varchar 职位描述 post_comment varchar 薪水 salary double 办公室 office int 部门编号 depart_id int #创建表 create table employee( id int not null unique auto_increment, name varchar(20) not null, sex enum('male','female') not null default 'male', #大部分是男的 age int(3) unsigned not null default 28, hire_date date not null, post varchar(50), post_comment varchar(100), salary double(15,2), office int, #一个部门一个屋子 depart_id int ); #查看表结构 mysql> desc employee; +--------------+-----------------------+------+-----+---------+----------------+ | Field | Type | Null | Key | Default | Extra | +--------------+-----------------------+------+-----+---------+----------------+ | id | int(11) | NO | PRI | NULL | auto_increment | | name | varchar(20) | NO | | NULL | | | sex | enum('male','female') | NO | | male | | | age | int(3) unsigned | NO | | 28 | | | hire_date | date | NO | | NULL | | | post | varchar(50) | YES | | NULL | | | post_comment | varchar(100) | YES | | NULL | | | salary | double(15,2) | YES | | NULL | | | office | int(11) | YES | | NULL | | | depart_id | int(11) | YES | | NULL | | +--------------+-----------------------+------+-----+---------+----------------+ #插入记录 #三个部门：教学，销售，运营 insert into employee(name,sex,age,hire_date,post,salary,office,depart_id) values ('egon','male',18,'20170301','老男孩驻沙河办事处外交大使',7300.33,401,1), #以下是教学部 ('alex','male',78,'20150302','teacher',1000000.31,401,1), ('wupeiqi','male',81,'20130305','teacher',8300,401,1), ('yuanhao','male',73,'20140701','teacher',3500,401,1), ('liwenzhou','male',28,'20121101','teacher',2100,401,1), ('jingliyang','female',18,'20110211','teacher',9000,401,1), ('jinxin','male',18,'19000301','teacher',30000,401,1), ('成龙','male',48,'20101111','teacher',10000,401,1), ('歪歪','female',48,'20150311','sale',3000.13,402,2),#以下是销售部门 ('丫丫','female',38,'20101101','sale',2000.35,402,2), ('丁丁','female',18,'20110312','sale',1000.37,402,2), ('星星','female',18,'20160513','sale',3000.29,402,2), ('格格','female',28,'20170127','sale',4000.33,402,2), ('张野','male',28,'20160311','operation',10000.13,403,3), #以下是运营部门 ('程咬金','male',18,'19970312','operation',20000,403,3), ('程咬银','female',18,'20130311','operation',19000,403,3), ('程咬铜','male',18,'20150411','operation',18000,403,3), ('程咬铁','female',18,'20140512','operation',17000,403,3) ; #ps：如果在windows系统中，插入中文字符，select的结果为空白，可以将所有字符编码统一设置成gbk 准备表和记录
SELECT * FROM emp AS t1 INNER JOIN ( SELECT post, max(hire_date) max_date FROM emp GROUP BY post ) AS t2 ON t1.post = t2.post WHERE t1.hire_date = t2.max_date; 答案一（链表）
mysql> select (select t2.name from emp as t2 where t2.post=t1.post order by hire_date desc limit 1) from emp as t1 group by post; +---------------------------------------------------------------------------------------+ | (select t2.name from emp as t2 where t2.post=t1.post order by hire_date desc limit 1) | +---------------------------------------------------------------------------------------+ | 张野 | | 格格 | | alex | | egon | +---------------------------------------------------------------------------------------+ rows in set (0.00 sec) mysql> select (select t2.id from emp as t2 where t2.post=t1.post order by hire_date desc limit 1) from emp as t1 group by post; +-------------------------------------------------------------------------------------+ | (select t2.id from emp as t2 where t2.post=t1.post order by hire_date desc limit 1) | +-------------------------------------------------------------------------------------+ | 14 | | 13 | | 2 | | 1 | +-------------------------------------------------------------------------------------+ rows in set (0.00 sec) #正确答案 mysql> select t3.name,t3.post,t3.hire_date from emp as t3 where id in (select (select id from emp as t2 where t2.post=t1.post order by hire_date desc limit 1) from emp as t1 group by post); +--------+-----------------------------------------+------------+ | name | post | hire_date | +--------+-----------------------------------------+------------+ | egon | 老男孩驻沙河办事处外交大使 | 2017-03-01 | | alex | teacher | 2015-03-02 | | 格格 | sale | 2017-01-27 | | 张野 | operation | 2016-03-11 | +--------+-----------------------------------------+------------+ rows in set (0.00 sec) 答案二（子查询）
答案一为正确答案，答案二中的limit 1有问题（每个部门可能有>1个为同一时间入职的新员工），我只是想用该例子来说明可以在select后使用子查询
可以基于上述方法解决：比如某网站在全国各个市都有站点，每个站点一条数据，想取每个省下最新的那一条市的网站质量信息
五 sql逻辑查询语句的执行顺序
init.sql文件内容
/* 数据导入： Navicat Premium Data Transfer Source Server : localhost Source Server Type : MySQL Source Server Version : 50624 Source Host : localhost Source Database : sqlexam Target Server Type : MySQL Target Server Version : 50624 File Encoding : utf-8 Date: 10/21/2016 06:46:46 AM */ SET NAMES utf8; SET FOREIGN_KEY_CHECKS = 0; -- ---------------------------- -- Table structure for `class` -- ---------------------------- DROP TABLE IF EXISTS `class`; CREATE TABLE `class` ( `cid` int(11) NOT NULL AUTO_INCREMENT, `caption` varchar(32) NOT NULL, PRIMARY KEY (`cid`) ) ENGINE=InnoDB AUTO_INCREMENT=5 DEFAULT CHARSET=utf8; -- ---------------------------- -- Records of `class` -- ---------------------------- BEGIN; INSERT INTO `class` VALUES ('1', '三年二班'), ('2', '三年三班'), ('3', '一年二班'), ('4', '二年九班'); COMMIT; -- ---------------------------- -- Table structure for `course` -- ---------------------------- DROP TABLE IF EXISTS `course`; CREATE TABLE `course` ( `cid` int(11) NOT NULL AUTO_INCREMENT, `cname` varchar(32) NOT NULL, `teacher_id` int(11) NOT NULL, PRIMARY KEY (`cid`), KEY `fk_course_teacher` (`teacher_id`), CONSTRAINT `fk_course_teacher` FOREIGN KEY (`teacher_id`) REFERENCES `teacher` (`tid`) ) ENGINE=InnoDB AUTO_INCREMENT=5 DEFAULT CHARSET=utf8; -- ---------------------------- -- Records of `course` -- ---------------------------- BEGIN; INSERT INTO `course` VALUES ('1', '生物', '1'), ('2', '物理', '2'), ('3', '体育', '3'), ('4', '美术', '2'); COMMIT; -- ---------------------------- -- Table structure for `score` -- ---------------------------- DROP TABLE IF EXISTS `score`; CREATE TABLE `score` ( `sid` int(11) NOT NULL AUTO_INCREMENT, `student_id` int(11) NOT NULL, `course_id` int(11) NOT NULL, `num` int(11) NOT NULL, PRIMARY KEY (`sid`), KEY `fk_score_student` (`student_id`), KEY `fk_score_course` (`course_id`), CONSTRAINT `fk_score_course` FOREIGN KEY (`course_id`) REFERENCES `course` (`cid`), CONSTRAINT `fk_score_student` FOREIGN KEY (`student_id`) REFERENCES `student` (`sid`) ) ENGINE=InnoDB AUTO_INCREMENT=53 DEFAULT CHARSET=utf8; -- ---------------------------- -- Records of `score` -- ---------------------------- BEGIN; INSERT INTO `score` VALUES ('1', '1', '1', '10'), ('2', '1', '2', '9'), ('5', '1', '4', '66'), ('6', '2', '1', '8'), ('8', '2', '3', '68'), ('9', '2', '4', '99'), ('10', '3', '1', '77'), ('11', '3', '2', '66'), ('12', '3', '3', '87'), ('13', '3', '4', '99'), ('14', '4', '1', '79'), ('15', '4', '2', '11'), ('16', '4', '3', '67'), ('17', '4', '4', '100'), ('18', '5', '1', '79'), ('19', '5', '2', '11'), ('20', '5', '3', '67'), ('21', '5', '4', '100'), ('22', '6', '1', '9'), ('23', '6', '2', '100'), ('24', '6', '3', '67'), ('25', '6', '4', '100'), ('26', '7', '1', '9'), ('27', '7', '2', '100'), ('28', '7', '3', '67'), ('29', '7', '4', '88'), ('30', '8', '1', '9'), ('31', '8', '2', '100'), ('32', '8', '3', '67'), ('33', '8', '4', '88'), ('34', '9', '1', '91'), ('35', '9', '2', '88'), ('36', '9', '3', '67'), ('37', '9', '4', '22'), ('38', '10', '1', '90'), ('39', '10', '2', '77'), ('40', '10', '3', '43'), ('41', '10', '4', '87'), ('42', '11', '1', '90'), ('43', '11', '2', '77'), ('44', '11', '3', '43'), ('45', '11', '4', '87'), ('46', '12', '1', '90'), ('47', '12', '2', '77'), ('48', '12', '3', '43'), ('49', '12', '4', '87'), ('52', '13', '3', '87'); COMMIT; -- ---------------------------- -- Table structure for `student` -- ---------------------------- DROP TABLE IF EXISTS `student`; CREATE TABLE `student` ( `sid` int(11) NOT NULL AUTO_INCREMENT, `gender` char(1) NOT NULL, `class_id` int(11) NOT NULL, `sname` varchar(32) NOT NULL, PRIMARY KEY (`sid`), KEY `fk_class` (`class_id`), CONSTRAINT `fk_class` FOREIGN KEY (`class_id`) REFERENCES `class` (`cid`) ) ENGINE=InnoDB AUTO_INCREMENT=17 DEFAULT CHARSET=utf8; -- ---------------------------- -- Records of `student` -- ---------------------------- BEGIN; INSERT INTO `student` VALUES ('1', '男', '1', '理解'), ('2', '女', '1', '钢蛋'), ('3', '男', '1', '张三'), ('4', '男', '1', '张一'), ('5', '女', '1', '张二'), ('6', '男', '1', '张四'), ('7', '女', '2', '铁锤'), ('8', '男', '2', '李三'), ('9', '男', '2', '李一'), ('10', '女', '2', '李二'), ('11', '男', '2', '李四'), ('12', '女', '3', '如花'), ('13', '男', '3', '刘三'), ('14', '男', '3', '刘一'), ('15', '女', '3', '刘二'), ('16', '男', '3', '刘四'); COMMIT; -- ---------------------------- -- Table structure for `teacher` -- ---------------------------- DROP TABLE IF EXISTS `teacher`; CREATE TABLE `teacher` ( `tid` int(11) NOT NULL AUTO_INCREMENT, `tname` varchar(32) NOT NULL, PRIMARY KEY (`tid`) ) ENGINE=InnoDB AUTO_INCREMENT=6 DEFAULT CHARSET=utf8; -- ---------------------------- -- Records of `teacher` -- ---------------------------- BEGIN; INSERT INTO `teacher` VALUES ('1', '张磊老师'), ('2', '李平老师'), ('3', '刘海燕老师'), ('4', '朱云海老师'), ('5', '李杰老师'); COMMIT; SET FOREIGN_KEY_CHECKS = 1;
从init.sql文件中导入数据
#准备表、记录 mysql> create database db1; mysql> use db1; mysql> source /root/init.sql #从init.sql文件中导入数据
！！！重中之重：练习之前务必搞清楚sql逻辑查询语句的执行顺序
链接：http://www.cnblogs.com/linhaifeng/articles/7372774.html
sql逻辑查询语句的执行顺序
1 SELECT语句关键字的定义顺序
SELECT DISTINCT <select_list> FROM <left_table> <join_type> JOIN <right_table> ON <join_condition> WHERE <where_condition> GROUP BY <group_by_list> HAVING <having_condition> ORDER BY <order_by_condition> LIMIT <limit_number>
2 SELECT语句关键字的执行顺序
(7) SELECT (8) DISTINCT <select_list> (1) FROM <left_table> (3) <join_type> JOIN <right_table> (2) ON <join_condition> (4) WHERE <where_condition> (5) GROUP BY <group_by_list> (6) HAVING <having_condition> (9) ORDER BY <order_by_condition> (10) LIMIT <limit_number>
3 准备表和数据
1. 新建一个测试数据库TestDB；
create database TestDB;
2.创建测试表table1和table2；
CREATE TABLE table1
(
customer_id VARCHAR(10) NOT NULL,
city VARCHAR(10) NOT NULL,
PRIMARY KEY(customer_id)
)ENGINE=INNODB DEFAULT CHARSET=UTF8;
CREATE TABLE table2
(
order_id INT NOT NULL auto_increment,
customer_id VARCHAR(10),
PRIMARY KEY(order_id)
)ENGINE=INNODB DEFAULT CHARSET=UTF8;
3.插入测试数据；
INSERT INTO table1(customer_id,city) VALUES('163','hangzhou');
INSERT INTO table1(customer_id,city) VALUES('9you','shanghai');
INSERT INTO table1(customer_id,city) VALUES('tx','hangzhou');
INSERT INTO table1(customer_id,city) VALUES('baidu','hangzhou');
INSERT INTO table2(customer_id) VALUES('163');
INSERT INTO table2(customer_id) VALUES('163');
INSERT INTO table2(customer_id) VALUES('9you');
INSERT INTO table2(customer_id) VALUES('9you');
INSERT INTO table2(customer_id) VALUES('9you');
INSERT INTO table2(customer_id) VALUES('tx');
INSERT INTO table2(customer_id) VALUES(NULL);
准备工作做完以后，table1和table2看起来应该像下面这样：
mysql> select * from table1; +-------------+----------+ | customer_id | city | +-------------+----------+ | 163 | hangzhou | | 9you | shanghai | | baidu | hangzhou | | tx | hangzhou | +-------------+----------+ 4 rows in set (0.00 sec) mysql> select * from table2; +----------+-------------+ | order_id | customer_id | +----------+-------------+ | 1 | 163 | | 2 | 163 | | 3 | 9you | | 4 | 9you | | 5 | 9you | | 6 | tx | | 7 | NULL | +----------+-------------+ 7 rows in set (0.00 sec)
4 准备SQL逻辑查询测试语句
#查询来自杭州，并且订单数少于2的客户。 SELECT a.customer_id, COUNT(b.order_id) as total_orders FROM table1 AS a LEFT JOIN table2 AS b ON a.customer_id = b.customer_id WHERE a.city = 'hangzhou' GROUP BY a.customer_id HAVING count(b.order_id) < 2 ORDER BY total_orders DESC;
5 执行顺序分析
在这些SQL语句的执行过程中，都会产生一个虚拟表，用来保存SQL语句的执行结果（这是重点），我现在就来跟踪这个虚拟表的变化，得到最终的查询结果的过程，来分析整个SQL逻辑查询的执行顺序和过程。
执行FROM语句
第一步，执行FROM语句。我们首先需要知道最开始从哪个表开始的，这就是FROM告诉我们的。现在有了<left_table>和<right_table>两个表，我们到底从哪个表开始，还是从两个表进行某种联系以后再开始呢？它们之间如何产生联系呢？——笛卡尔积
关于什么是笛卡尔积，请自行Google补脑。经过FROM语句对两个表执行笛卡尔积，会得到一个虚拟表，暂且叫VT1（vitual table 1），内容如下：
+-------------+----------+----------+-------------+ | customer_id | city | order_id | customer_id | +-------------+----------+----------+-------------+ | 163 | hangzhou | 1 | 163 | | 9you | shanghai | 1 | 163 | | baidu | hangzhou | 1 | 163 | | tx | hangzhou | 1 | 163 | | 163 | hangzhou | 2 | 163 | | 9you | shanghai | 2 | 163 | | baidu | hangzhou | 2 | 163 | | tx | hangzhou | 2 | 163 | | 163 | hangzhou | 3 | 9you | | 9you | shanghai | 3 | 9you | | baidu | hangzhou | 3 | 9you | | tx | hangzhou | 3 | 9you | | 163 | hangzhou | 4 | 9you | | 9you | shanghai | 4 | 9you | | baidu | hangzhou | 4 | 9you | | tx | hangzhou | 4 | 9you | | 163 | hangzhou | 5 | 9you | | 9you | shanghai | 5 | 9you | | baidu | hangzhou | 5 | 9you | | tx | hangzhou | 5 | 9you | | 163 | hangzhou | 6 | tx | | 9you | shanghai | 6 | tx | | baidu | hangzhou | 6 | tx | | tx | hangzhou | 6 | tx | | 163 | hangzhou | 7 | NULL | | 9you | shanghai | 7 | NULL | | baidu | hangzhou | 7 | NULL | | tx | hangzhou | 7 | NULL | +-------------+----------+----------+-------------+
总共有28（table1的记录条数 * table2的记录条数）条记录。这就是VT1的结果，接下来的操作就在VT1的基础上进行。
执行ON过滤
执行完笛卡尔积以后，接着就进行ON a.customer_id = b.customer_id条件过滤，根据ON中指定的条件，去掉那些不符合条件的数据，得到VT2表，内容如下：
+-------------+----------+----------+-------------+ | customer_id | city | order_id | customer_id | +-------------+----------+----------+-------------+ | 163 | hangzhou | 1 | 163 | | 163 | hangzhou | 2 | 163 | | 9you | shanghai | 3 | 9you | | 9you | shanghai | 4 | 9you | | 9you | shanghai | 5 | 9you | | tx | hangzhou | 6 | tx | +-------------+----------+----------+-------------+
VT2就是经过ON条件筛选以后得到的有用数据，而接下来的操作将在VT2的基础上继续进行。
添加外部行
这一步只有在连接类型为OUTER JOIN时才发生，如LEFT OUTER JOIN、RIGHT OUTER JOIN和FULL OUTER JOIN。在大多数的时候，我们都是会省略掉OUTER关键字的，但OUTER表示的就是外部行的概念。
LEFT OUTER JOIN把左表记为保留表，得到的结果为：
+-------------+----------+----------+-------------+ | customer_id | city | order_id | customer_id | +-------------+----------+----------+-------------+ | 163 | hangzhou | 1 | 163 | | 163 | hangzhou | 2 | 163 | | 9you | shanghai | 3 | 9you | | 9you | shanghai | 4 | 9you | | 9you | shanghai | 5 | 9you | | tx | hangzhou | 6 | tx | | baidu | hangzhou | NULL | NULL | +-------------+----------+----------+-------------+
RIGHT OUTER JOIN把右表记为保留表，得到的结果为：
+-------------+----------+----------+-------------+ | customer_id | city | order_id | customer_id | +-------------+----------+----------+-------------+ | 163 | hangzhou | 1 | 163 | | 163 | hangzhou | 2 | 163 | | 9you | shanghai | 3 | 9you | | 9you | shanghai | 4 | 9you | | 9you | shanghai | 5 | 9you | | tx | hangzhou | 6 | tx | | NULL | NULL | 7 | NULL | +-------------+----------+----------+-------------+
FULL OUTER JOIN把左右表都作为保留表，得到的结果为：
+-------------+----------+----------+-------------+ | customer_id | city | order_id | customer_id | +-------------+----------+----------+-------------+ | 163 | hangzhou | 1 | 163 | | 163 | hangzhou | 2 | 163 | | 9you | shanghai | 3 | 9you | | 9you | shanghai | 4 | 9you | | 9you | shanghai | 5 | 9you | | tx | hangzhou | 6 | tx | | baidu | hangzhou | NULL | NULL | | NULL | NULL | 7 | NULL | +-------------+----------+----------+-------------+
添加外部行的工作就是在VT2表的基础上添加保留表中被过滤条件过滤掉的数据，非保留表中的数据被赋予NULL值，最后生成虚拟表VT3。
由于我在准备的测试SQL查询逻辑语句中使用的是LEFT JOIN，过滤掉了以下这条数据：
| baidu | hangzhou | NULL | NULL |
现在就把这条数据添加到VT2表中，得到的VT3表如下：
+-------------+----------+----------+-------------+ | customer_id | city | order_id | customer_id | +-------------+----------+----------+-------------+ | 163 | hangzhou | 1 | 163 | | 163 | hangzhou | 2 | 163 | | 9you | shanghai | 3 | 9you | | 9you | shanghai | 4 | 9you | | 9you | shanghai | 5 | 9you | | tx | hangzhou | 6 | tx | | baidu | hangzhou | NULL | NULL | +-------------+----------+----------+-------------+
接下来的操作都会在该VT3表上进行。
执行WHERE过滤
对添加外部行得到的VT3进行WHERE过滤，只有符合<where_condition>的记录才会输出到虚拟表VT4中。当我们执行WHERE a.city = 'hangzhou'的时候，就会得到以下内容，并存在虚拟表VT4中：
+-------------+----------+----------+-------------+ | customer_id | city | order_id | customer_id | +-------------+----------+----------+-------------+ | 163 | hangzhou | 1 | 163 | | 163 | hangzhou | 2 | 163 | | tx | hangzhou | 6 | tx | | baidu | hangzhou | NULL | NULL | +-------------+----------+----------+-------------+
但是在使用WHERE子句时，需要注意以下两点：
- 由于数据还没有分组，因此现在还不能在WHERE过滤器中使用
where_condition=MIN(col)这类对分组统计的过滤； - 由于还没有进行列的选取操作，因此在SELECT中使用列的别名也是不被允许的，如：
SELECT city as c FROM t WHERE c='shanghai';是不允许出现的。
执行GROUP BY分组
GROU BY子句主要是对使用WHERE子句得到的虚拟表进行分组操作。我们执行测试语句中的GROUP BY a.customer_id，就会得到以下内容(默认只显示组内第一条)：
+-------------+----------+----------+-------------+ | customer_id | city | order_id | customer_id | +-------------+----------+----------+-------------+ | 163 | hangzhou | 1 | 163 | | baidu | hangzhou | NULL | NULL | | tx | hangzhou | 6 | tx | +-------------+----------+----------+-------------+
得到的内容会存入虚拟表VT5中，此时，我们就得到了一个VT5虚拟表，接下来的操作都会在该表上完成。
执行HAVING过滤
HAVING子句主要和GROUP BY子句配合使用，对分组得到的VT5虚拟表进行条件过滤。当我执行测试语句中的HAVING count(b.order_id) < 2时，将得到以下内容：
+-------------+----------+----------+-------------+ | customer_id | city | order_id | customer_id | +-------------+----------+----------+-------------+ | baidu | hangzhou | NULL | NULL | | tx | hangzhou | 6 | tx | +-------------+----------+----------+-------------+
这就是虚拟表VT6。
SELECT列表
现在才会执行到SELECT子句，不要以为SELECT子句被写在第一行，就是第一个被执行的。
我们执行测试语句中的SELECT a.customer_id, COUNT(b.order_id) as total_orders，从虚拟表VT6中选择出我们需要的内容。我们将得到以下内容：
+-------------+--------------+ | customer_id | total_orders | +-------------+--------------+ | baidu | 0 | | tx | 1 | +-------------+--------------+
还没有完，这只是虚拟表VT7。
执行DISTINCT子句
如果在查询中指定了DISTINCT子句，则会创建一张内存临时表（如果内存放不下，就需要存放在硬盘了）。这张临时表的表结构和上一步产生的虚拟表VT7是一样的，不同的是对进行DISTINCT操作的列增加了一个唯一索引，以此来除重复数据。
由于我的测试SQL语句中并没有使用DISTINCT，所以，在该查询中，这一步不会生成一个虚拟表。
执行ORDER BY子句
对虚拟表中的内容按照指定的列进行排序，然后返回一个新的虚拟表，我们执行测试SQL语句中的ORDER BY total_orders DESC，就会得到以下内容：
+-------------+--------------+ | customer_id | total_orders | +-------------+--------------+ | tx | 1 | | baidu | 0 | +-------------+--------------+
可以看到这是对total_orders列进行降序排列的。上述结果会存储在VT8中。
执行LIMIT子句
LIMIT子句从上一步得到的VT8虚拟表中选出从指定位置开始的指定行数据。对于没有应用ORDER BY的LIMIT子句，得到的结果同样是无序的，所以，很多时候，我们都会看到LIMIT子句会和ORDER BY子句一起使用。
MySQL数据库的LIMIT支持如下形式的选择：
LIMIT n, m
表示从第n条记录开始选择m条记录。而很多开发人员喜欢使用该语句来解决分页问题。对于小数据，使用LIMIT子句没有任何问题，当数据量非常大的时候，使用LIMIT n, m是非常低效的。因为LIMIT的机制是每次都是从头开始扫描，如果需要从第60万行开始，读取3条数据，就需要先扫描定位到60万行，然后再进行读取，而扫描的过程是一个非常低效的过程。所以，对于大数据处理时，是非常有必要在应用层建立一定的缓存机制（现在的大数据处理，大都使用缓存）
