目录

外键约束

概念

特点

操作

多表联合查询

概念

操作

多表操作总结

外键约束

概念

特点

定义一个外键时，需要遵守下列规则：

主表必须已经存在于数据库中，或者是当前正在创建的表。

必须为主表定义主键。

主键不能包含空值，但允许在外键中出现空值。也就是说，只要外键的每个非空值出现在指定的主键中，这 个外键的内容就是正确的。

在主表的表名后面指定列名或列名的组合。这个列或列的组合必须是主表的主键或候选键。

外键中列的数目必须和主表的主键中列的数目相同。

外键中列的数据类型必须和主表主键中对应列的数据类型相同。

操作

• 建立外键约束

create database  mydb3;
use mydb3;
create table if not exists dep
(
pid int primary key,
name varchar(20)
);
create table if not exists per
(
id int primary key,
name varchar(20),
age int,
depid int,
constraint fok foreign key(depid) references dep(pid)
);create table if not exists dep3
(
pid int primary key,
name varchar(20)
);
create table if not exists per3
(
id int primary key,
name varchar(20),
age int,
depid int
);
alter table per3 add constraint fok3 foreign key(depid) references dep3(pid);

•  数据插入

必须先给主表添加数据，且从表外键列的值必须依赖于主表的主键列

insert into dep3 values('1001','研发部');
insert into dep3 values('1002','销售部');
insert into dep3 values('1003','财务部');
insert into dep3 values('1004','人事部');-- 给per3表添加数据
insert into per3 values('1','乔峰',20, '1001');
insert into per3 values('2','段誉',21, '1001');
insert into per3 values('3','虚竹',23, '1001');
insert into per3 values('4','阿紫',18, '1001');
insert into per3 values('5','扫地僧',85, '1002');
insert into per3 values('6','李秋水',33, '1002');
insert into per3 values('7','鸠摩智',50, '1002'); 
insert into per3 values('8','天山童姥',60, '1003');
insert into per3 values('9','慕容博',58, '1003');

• 数据删除

主表数据被从表依赖时不能删除，否则可以删除；从表的数据可以随便删除。

如下，第一句和第二句执行成功，第三句执行失败

delete from per3 where depid=1003;
delete from dep3 where pid=1004;
delete from dep3 where pid=1002;

• 删除外键约束

语法：alter table 从表 drop foreign key 关键词名;

alter table per3 drop foreign key fok3;

多表联合查询

概念

操作

• 交叉连接查询

select * from dept,emp;

• 内连接查询

注释；上面是隐式内连接，下面是显式内连接 

select * from dept,emp where dept.deptno=emp.dept_id;
select * from dept join emp on dept.deptno=emp.dept_id;select * from dept join emp on dept.deptno=emp.dept_id and name='研发部';
select * from dept join emp on dept.deptno=emp.dept_id and name='研发部';select * from dept join emp on dept.deptno=emp.dept_id and (name='研发部' or name='销售部');
select * from dept join emp on dept.deptno=emp.dept_id and (name='研发部' or name ='销售部');
select * from dept join emp on dept.deptno=emp.dept_id and name in ('研发部','销售部');select a.name,a.deptno,count(*) from dept a join emp on a.deptno=emp.dept_id group by dept_id;
select a.name,a.deptno,count(*) total from dept a join emp on a.deptno=emp.dept_id group by dept_id having total >=3 order by total desc;

• 外连接查询

若是对应的外表没有数据就补NULL

select * from dept a left join emp b on a.deptno=b.dept_id;
select * from dept a right join emp b on a.deptno=b.dept_id;
-- select * from dept a full join emp b on a.deptno=b.dept_id; --不能执行
-- 用下面的方法代替上面的full join 
select * from dept a left join emp b on a.deptno=b.dept_id union select * from dept a right join emp b on a.deptno=b.dept_id;
-- 对比union all，发现union all没有去重过滤
select * from dept a left join emp b on a.deptno=b.dept_id union all select * from dept a right join emp b on a.deptno=b.dept_id;

• 子查询

select * from emp where age<(select avg(age) from emp);
select * from emp a where a.dept_id in (select deptno from dept where name in ('研发部','销售部'));
-- 对比关联查询和子查询如下
select * from emp a join dept b on a.dept_id=b.deptno and (b.name='研发部' and age<30);
select * from (select * from dept where name='研发部') a join (select * from emp where age<30) b on b.dept_id=a.deptno;

• 子查询关键字

all关键字的用法

select * from emp where age>all(select age from emp where dept_id='1003');
select * from emp a where a.dept_id!=all(select deptno from dept);

any（some）关键字的用法

select * from emp where age>any(select age from emp where dept_id='1003') and dept_id!='1003';

in关键字的用法

select ename,eid from emp where dept_id in (select deptno from dept where name in ('研发部','销售部'));

 exists关键字的用法

select * from emp a where a.age<30;
select * from emp a where exists(select * from emp where a.age<30);select * from emp a where a.dept_id in (select deptno from dept b);
select * from emp a where exists (select * from dept b where a.dept_id = b.deptno);

• 自关联查询

多表操作总结