ROLLUP和CUBE 用法
Oracle的GROUP BY语句除了最基本的语法外，还支持ROLLUP和CUBE语句。如果是Group by ROLLUP(A, B, C)的话，首先会对(A、B、C)进行GROUP BY，然后对(A、B)进行GROUP BY，然后是(A)进行GROUP BY，最后对全表进行GROUP BY操作。
如果是GROUP BY CUBE(A, B, C)，则首先会对(A、B、C)进行GROUP BY，然后依次巧困是(A、B)，(A、C)，(A)，(B、C)，(B)，(C)，最后对全表进行GROUP BY操作。 grouping_id()可以美化效果。除了使用GROUPING函数，还可以使用GROUPING_ID来标识GROUP BY的结果。
也可以 Group by Rollup(A,(B,C)) ，Group by A Rollup(B,C)，…… 这样任意按自己想要的形式结合统计数据，非常方便。
Rollup()：分组函数可以理解散弯为group by的精简模式，具体分组模式如下：
Rollup(a,b,c): (a,b,c),(a,b),(a),(全表)
Cube():分组函数也是以group by为基础，具体分组模式如下：
cube(a,b,c):(a,b,c),(a,b),(a,c),(b,c),(a),(b),(c),(全表)
下面准备数据比较一下两个函数的不同：
1、准备数据：
2、使用rollup函数查询
select group_id,job,name,sum(salary) from GROUP_TEST group by rollup(group_id,job,name)；
3、使用cube函数：
select group_id,job,name,sum(salary) from GROUP_TEST group by cube(group_id,job,name)
4、对孝掘念比：从最后查询出来的数据条数就差了好多，下面看一下将两个函数从转化成对应的group函数语句：
rollup函数：
select group_id,job,name,sum(salary) from GROUP_TEST group by rollup(group_id,job,name);
等价于：
select group_id,job,name,sum(salary) from GROUP_TEST group by group_id,job,name
union all
select group_id,job,null,sum(salary) from GROUP_TEST group by group_id,job
union all
select group_id,null,null,sum(salary) from GROUP_TEST group by group_id
union all
select null,null,null,sum(salary) from GROUP_TEST
cube函数：
select group_id,job,name,sum(salary) from GROUP_TEST group by cube(group_id,job,name) ;
等价于：
select group_id,job,name,sum(salary) from GROUP_TEST group by group_id,job,name
union all
select group_id,job,null,sum(salary) from GROUP_TEST group by group_id,job
union all
select group_id,null,name,sum(salary) from GROUP_TEST group by group_id,name
union all
select group_id,null,null,sum(salary) from GROUP_TEST group by group_id
union all
select null,job,name,sum(salary) from GROUP_TEST group by job,name
union all
select null,job,null,sum(salary) from GROUP_TEST group by job
union all
select null,null,name,sum(salary) from GROUP_TEST group by name
union all
select null,null,null,sum(salary) from GROUP_TEST
5、由此可见两个函数对于汇总统计来说要比普通函数好用的多，另外还有一个配套使用的函数
grouping(**):当**字段为null的时候值为1，当字段**非null的时候值为0；
select grouping(group_id),job,name,sum(salary) from GROUP_TEST group by rollup(group_id,job,name);
6、添加一列用来直观的显示所有的汇总字段：
select group_id,job,name,
case when name is null and nvl(group_id,0)=0 and job is null then '全表聚合'
when name is null and nvl(group_id,0)=0 and job is not null then 'JOB聚合'
when name is null and grouping(group_id)=0 and job is null then 'GROUPID聚合'
when name is not null and nvl(group_id,0)=0 and job is null then 'Name聚合'
when name is not null and grouping(group_id)=0 and job is null then 'GROPName聚合'
when name is not null and grouping(group_id)=1 and job is not null then 'JOBName聚合'
when name is null and grouping(group_id)=0 and job is not null then 'GROUPJOB聚合'
else
'三列汇总' end ,
sum(salary) from GROUP_TEST group by cube(group_id,job,name) ;
