oracle中join学习个人笔记

3 joins
3.2连接条件
using
相同的列连接可用using子句，且select列中不能对条件列限定表名
using子句可使用多个列,如using(a,b)
SELECT location_id, department.name, location.regional_group
FROM department JOIN location
USING (location_id);

natural join
无需使用using子句
select location_id,department.name,location.regional_group
from location natural join department
会自动匹配两个表中所有相同的列，影响输出

3.3连接类型
cross join 笛卡尔积
不能带条件,如on子句

inner join 内连接
默认的连接，inner可省略

outer join 外连接
FROM table1 { LEFT | RIGHT | FULL } [OUTER] JOIN table2
left:返回table1中的所有记录，table2对应列显示空
right:返回table2中的所有记录，table1对应列显示空
full: 不匹配的记录全部显示
out:默认可缺省,oracle会自动在left,right,full后增加out

left out join
select d.dept_id,d.name,l.regional_group
from department d left outer join location l
on d.location_id=l.location_id
早期的写法
select d.dept_id,d.name,l.regional_group
from department d,location l
where d.location_id=l.location_id(+)
返回department表里所有记录

right out join
select d.dept_id,d.name,l.regional_group
from department d right outer join location l
on d.location_id=l.location_id
早期的写法
select d.dept_id,d.name,l.regional_group
from department d,location l
where d.location_id(+)=l.location_id
返回location表里所有记录

full out join
两张表不匹配的记录都显示

相等连接与不等连接
不等连接示例
select p.name part_name,c.inv_class inv_class
from part p join inventory_class c
on p.unit_cost between c.low_cost and c.high_cost

self join
自联接
Self outer joins
select e.lname employee, m.lname manager
from employee e left outer join employee m
on e.manager_emp_id = m.emp_id

partition outer joins
oracle10g的新特性,与lag,lap查询有关系
select nvl(ee.emp_id, 7782), m.year, m.month, nvl(ee.expense_claim, 0)
from (select * from months where year = 2002) m
left outer join (select * from employee_expense where emp_id = 7782) ee
on m.month = ee.month
and m.year = ee.year
order by m.month

可用partition outer joins实现
SELECT ee.emp_id, m.year, m.month, NVL(ee.expense_claim, 0)
FROM (SELECT * FROM months WHERE year = 2002) m
LEFT OUTER JOIN employee_expense ee PARTITION BY(ee.emp_id)
ON m.year = ee.year
AND m.month = ee.month
ORDER BY ee.emp_id, m.month;
原理
首先对employee_expense按emp_id分组，一个值一组
然后每组都与months表联接，会有多个外联接操作

3.4joins and subqueries
联接与子查询
SELECT supplier_id, name
FROM supplier s
WHERE EXISTS (SELECT *
FROM part p
WHERE p.inventory_qty < 10
AND p.supplier_id = s.supplier_id);
可用以下表联接查询实现
SELECT s.supplier_id, s.name
FROM supplier s
JOIN part p
ON p.supplier_id = s.supplier_id
WHERE p.inventory_qty < 10;

3.5对联接视图执行DML
聚合函数 Aggregate functions, such as AVG, COUNT, MAX, MIN, SUM
分析函数 CUME_DIST
表联接视图如果有以下情况不能执行DML
层次查询语句，如start with 或 connect by
group by 或 order by
Model查询?
集合操作符，如union,intersect,minus
聚合函数
分析函数
select部分有子查询或关联查询
distinct
有read only选项
有rownum伪列

键保留表(key-preserved table)不是表的特性，是对联接视图中的表而言
对此类视图的dml是对上述表的dml

转自http://space.itpub.net/777981/viewspace-674911