영우 글로벌 러닝 4차혁명 교육과정의 "방형욱" 강사님과 함께 공부한 내용입니다.
# Rollup
- 컬럼의 개수가 n개면 n+1 가지 종류의 결과 생성
- 컬럼의 나열 순서가 중요함
select deptno, job, sum(sal)
from emp
group by deptno, job
select deptno, job, sum(sal)
from emp
group by ROLLUP(deptno, job);
select deptno, job, sum(sal)
from emp
group by GROUPING SETS((deptno, job),
(deptno), ());
cf. select deptno, decode(deptno, null,
'TOTAL', decode(job, null, 'DEPT_SUM', job)) as job, sum(sal)
from emp
group by ROLLUP(deptno, job);
# Cube
- 컬럼의 개수가 n개면 2^n 가지 종류의 결과 생성
- 컬럼의 나열 순서가 중요하지 않음
select deptno, job, sum(sal)
from emp
group by CUBE(deptno, job);
select deptno, job, sum(sal)
from emp
group by GROUPING SETS((deptno, job),
(deptno), (job), ());
# 집계 기준 컬럼이 3개일 경우
drop table emp2 purge;
create table emp2
as
select
empno, ename, sal, job, deptno, decode(mod(empno, 2), 0, 'M', 'W') as
gender
from emp e
union all
select
empno, ename, sal+100, job, deptno, decode(mod(empno, 2), 0, 'W', 'M')
as gender
from emp e;
select * from emp2;
--
select deptno, job, gender, sum(sal)
from emp2
group by ROLLUP(deptno, job, gender)
order by 1, 2, 3;
select deptno, job, gender, sum(sal)
from emp2
group by GROUPING SETS((deptno, job, gender),
(deptno, job), (deptno), ())
order by 1, 2, 3;
--
select deptno, job, gender, sum(sal)
from emp2
group by CUBE(deptno, job, gender)
order by 1, 2, 3;
select deptno, job, gender, sum(sal)
from emp2
group by GROUPING SETS((deptno, job, gender),
(deptno, job), (deptno, gender), (job, gender), (deptno), (job), (gender), ())
order by 1, 2, 3;
select deptno, job, gender, sum(sal)
from emp2
group by GROUPING SETS((deptno, job, gender),
(job, gender), ())
order by 1, 2, 3;
# GROUPING 함수
drop table emp3 purge;
create table emp3 as select * from emp;
update emp3 set job = null where rownum = 1;
select * from emp3;
=> grouping 함수를 사용하지 않을 경우 아래와 같이 질의 결과가
이상해집니다.
select deptno, job, sum(sal) as sum_sal
from emp3
group by ROLLUP(deptno, job);
select *
from (select deptno, job, sum(sal) as sum_sal
from emp3
group by ROLLUP(deptno, job))
where deptno is not null
and job is null;
=> grouping 함수를 사용하면 이렇게 달라집니다.
select deptno, job, sum(sal) as sum_sal,
grouping(deptno) as g_deptno,
grouping(job) as g_job
from emp3
group by ROLLUP(deptno, job);
select deptno, job, sum_sal
from (select deptno, job, sum(sal) as
sum_sal,
grouping(deptno) as g_deptno,
grouping(job) as g_job
from emp3
group by ROLLUP(deptno, job))
where g_deptno = 0
and
g_job = 1;
select deptno, job, sum_sal
from (select deptno, job, sum(sal) as
sum_sal,
grouping(deptno) as g_deptno,
grouping(job) as g_job
from emp3
group by ROLLUP(deptno, job))
where g_deptno = 0
and
g_job in (0, 1);
# Composite column
select deptno, job, gender, sum(sal)
from emp2
group by deptno, job, gender;
select deptno, job, gender, sum(sal)
from emp2
group by rollup(deptno, job, gender);
--> deptno, job, gender
--> deptno, job
--> deptno
--> ()
select deptno, job, gender, sum(sal)
from emp2
group by rollup(deptno, (job, gender));
--> deptno, job, gender
--> deptno
--> ()
select deptno, job, gender, sum(sal)
from emp2
group by rollup((deptno, job), gender);
--> deptno, job, gender
--> deptno, job
--> ()
select deptno, job, gender, sum(sal)
from emp2
group by rollup((deptno, job, gender));
--> deptno, job, gender
--> ()
# Concatenated
Groupings
select deptno, job, gender, sum(sal)
from emp2
group by deptno, rollup(job), cube(gender);
deptno job
gender
() ()
--> deptno, job, gender
--> deptno, job
--> deptno, gender
--> deptno
select deptno, job, gender, sum(sal)
from emp2
group by GROUPING SETS((deptno, job, gender),
(deptno, job), (deptno, gender), (deptno));
