group by

在介绍GROUP BY 和 HAVING 子句前，我们必需先讲讲sql语言中一种特殊的函数：聚合函数，例如SUM, COUNT, MAX, AVG等。这些函数和其它函数的根本区别就是它们一般作用在多条记录上。

SELECT SUM(population) FROM bbc

　　这里的SUM作用在所有返回记录的population字段上，结果就是该查询只返回一个结果，即所有国家的总人口数。

　　通过使用GROUP BY 子句，可以让SUM 和 COUNT 这些函数对属于一组的数据起作用。当你指定 GROUP BY region 时，属于同一个region（地区）的一组数据将只能返回一行值，也就是说，表中所有除region（地区）外的字段，只能通过 SUM, COUNT等聚合函数运算后返回一个值。

　　HAVING子句可以让我们筛选成组后的各组数据，WHERE子句在聚合前先筛选记录．也就是说作用在GROUP BY 子句和HAVING子句前．
而 HAVING子句在聚合后对组记录进行筛选。

　　让我们还是通过具体的实例来理解GROUP BY 和 HAVING 子句，还采用第三节介绍的bbc表。

　　SQL实例：

　　一、显示每个地区的总人口数和总面积：

SELECT region, SUM(population), SUM(area) FROM bbc GROUP BY region

　　先以region把返回记录分成多个组，这就是GROUP BY的字面含义。分完组后，然后用聚合函数对每组中的不同字段（一或多条记录）作运算。

　　二、 显示每个地区的总人口数和总面积．仅显示那些面积超过1000000的地区。

SELECT region, SUM(population), SUM(area) FROM bbc GROUP BY region HAVING SUM(area)>1000000

　　在这里，我们不能用where来筛选超过1000000的地区，因为表中不存在这样一条记录。

　　相反，HAVING子句可以让我们筛选成组后的各组数据．

group by 的增强、高级子查询、DML和DDL语句扩展

组函数
avg,sum,count,min,max

select id,sum(age) from users group by id having sum(age)>40;

超级聚合
rollup,cube
rollup从右往左再聚合
select id,num,sum(age) from users group by rollup(id,num)
cube从左往右取
select id,num,sum(age) from users group by cube(id,num)

grouping函数,返回的值，1肯定，0否定

select id,sum(age),grouping(id) from users;

grouping sets
分别求统计，再union
select id,num,sum(age) from users group by grouping sets((id),(num))

select id,num,sum(age) from users group by rollup((id,num));

select id,num,sum(age) from users group by id,rollup(num);

高级子查询

inner query ,outer query

select name from users where age >(select age from users where id =322);

成对的进行比较

select student_id,teacher_id,class_id from students
where (teacher_id,class_id) in (select teacher_id,class_id from students where student_id in (222,332))

在from中使用子查询,注意要加上别名

select a.id,a.name,b.type from students a,(select type from teachers) b where a.teacher_id=b.id;

标量子查询

精确的从一行中返回一列,独立的

关联子查询
循环的方式加以执行，内查询依赖于外查询
select name,age from users a
where age >(select avg(age) from users b where b.id=a.id )

exists,not exists

select * from students t1 where exists (select * from student_details t2 where t1.id=t2.id and t2.detail_id=22);

关联更新
用另外一张表更新当前表

update test1 t2 set age= (select age from test2 t2 where t1.id=t2.id)

关联删除

delete from test1 t1 where id in (select id from test2 t2 where t1.id=t2.id and age >20))

分层结构
select id,name from student start with id=292 connect by prior id=c_id;

select level,id,name from student start with id=292 connect by prior id=c_id;

DML和DDL语句扩展

DML:INSERT,DELETE,UPDATE

多表插入
无条件insert
有条件all insert
有条件first insert
选择insert

insert [all] [conditional_insert_clause]
[insert_into_clause values_clause] (subquery)

insert all into test1(id,name) values(id,name) into test2(id,address) values(id,address)
select id,name,address from test3 where id>23;

有条件的插入
insert all
when name like 't%' then
into test1(id,name) values(id,name)
when name like 'h%' then
into test2(id,address) values(id,address)
select id,name,address from test3 where id>23;

旋转插入
假设表sales_source_date的表结构为,employee_id,week_id,sales_mon,sales_tue,sales_wed,sales_thu,sales_fri
我想把这些记录保存在sales_info表中，表结构为employee_id,week,sales
insert all
into sales_info values (employee_id,week_id,sales_mon)
into sales_info values (employee_id,week_id,sales_tue)
into sales_info values (employee_id,week_id,sales_wed)
into sales_info values (employee_id,week_id,sales_thu)
into sales_info values (employee_id,week_id,sales_fri)
select employee_id,week_id,sales_mon,sales_tue,sales_wed,sales_thu,sales_fri from sales_source_date;

创建表时创建索引
create table test_index_table
(id int not null primary key using index(create index index1 on test_index_table(id)),