之前写过一篇行转列的文章：Oracle 简单的列转行
SQL> select * from v$version;
BANNER
--------------------------------------------------------------------------------
Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - 64bit Production
PL/SQL Release 11.2.0.1.0 - Production
CORE 11.2.0.1.0 Production
TNS for Linux: Version 11.2.0.1.0 - Production
NLSRTL Version 11.2.0.1.0 - Production
需求是，统计每个部门不同工种的薪水总和。
SQL> select deptno,ename,job,sal from emp;
DEPTNO ENAME JOB SAL
---------- ---------- --------- ----------
20 SMITH CLERK 800
30 ALLEN SALESMAN 1600
30 WARD SALESMAN 1250
20 JONES MANAGER 2975
30 MARTIN SALESMAN 1250
30 BLAKE MANAGER 2850
10 CLARK MANAGER 2450
20 SCOTT ANALYST 3000
10 KING PRESIDENT 5000
30 TURNER SALESMAN 1500
20 ADAMS CLERK 1100
30 JAMES CLERK 950
20 FORD ANALYST 3000
10 MILLER CLERK 1300
已选择14行。
SQL> select deptno,
nvl(sum(decode(job, ‘MANAGER‘, sal)), 0) "s_MANAGER",
nvl(sum(decode(job, ‘ANALYST‘, sal)), 0) "s_ANALYST",
nvl(sum(decode(job, ‘CLERK‘, sal)), 0) "s_CLERK",
nvl(sum(decode(job, ‘PRESIDENT‘, sal)), 0) "s_PRESIDENT",
nvl(sum(decode(job, ‘SALESMAN‘, sal)), 0) "s_SALESMAN"
from emp
group by deptno;
DEPTNO s_MANAGER s_ANALYST s_CLERK s_PRESIDENT s_SALESMAN
---------- ---------- ---------- ---------- ----------- ----------
30 2850 0 950 0 5600
20 2975 6000 1900 0 0
10 2450 0 1300 5000 0
用pivot会更简洁
SQL> with p as (select deptno,job,sal from emp)
SELECT * FROM p pivot ( SUM(sal)
FOR job IN (‘MANAGER‘ as "s_MANAGER",
‘ANALYST‘ as "s_ANALYST",
‘CLERK‘ as "s_CLERK",
‘PRESIDENT‘ as "s_PRESIDENT",
‘SALESMAN‘ as "s_SALESMAN" ));
DEPTNO s_MANAGER s_ANALYST s_CLERK s_PRESIDENT s_SALESMAN
---------- ---------- ---------- ---------- ----------- ----------
30 2850 950 5600
20 2975 6000 1900
10 2450 1300 5000