listagg小记录

    listagg的作用是将分组范围内的所有行特定列的记录加以合并成行。函数签名中的measure_expr为分组中每个列的表达式，而delimiter为合并分割符。如果delimiter不设置的话，就表示无分割符。

    中间within group后面的order_by_clause表示的是进行合并中要遵守的排序顺序。而后面的over子句表明listagg是具有分析函数analyze funcation特性的。具体采用listagg有三个场景。

1、对于无分组  

SQL> select * from emp where deptno=30; 

EMPNO ENAME     JOB        MGR HIREDATE         SAL     COMM DEPTNO

----- ---------- --------- ----- ----------- --------- --------- ------

 7499 ALLEN     SALESMAN  7698 1981-2-20    1600.00   300.00    30

 7521 WARD      SALESMAN  7698 1981-2-22    1250.00   500.00    30

 7654 MARTIN    SALESMAN  7698 1981-9-28    1250.00  1400.00    30

 7698 BLAKE     MANAGER   7839 1981-5-1     2850.00              30

 7844 TURNER    SALESMAN  7698 1981-9-8     1500.00     0.00    30

 7900 JAMES     CLERK     7698 1981-12-3     950.00              30 

6 rows selected

 

--按照empno进行排序

SQL> select listagg(ename,' , ') within group (order byempno) from emp where deptno=30; 

LISTAGG(ENAME,',')WITHINGROUP(

------------------------------------------------------------

ALLEN , WARD , MARTIN , BLAKE , TURNER , JAMES

 

 

2、    在有分组条件下的listagg使用

SQL> select deptno, listagg(ename,' ,') within group (order by empno) from emp group by deptno; 

DEPTNO LISTAGG(ENAME,',')WITHINGROUP(

------ -------------------------------------

   10 CLARK ,KING ,MILLER

   20 SMITH ,JONES ,SCOTT ,ADAMS ,FORD

   30 ALLEN ,WARD ,MARTIN ,BLAKE ,TURNER ,JAMES 

 

 3、 使用over分组情况 

 

SQL> select deptno, ename, listagg(ename, ' , ') within group (order by empno) over (partition by deptno) as emp_list  from emp

 2 where hiredate<=add_months(sysdate,-10*12);

DEPTNO ENAME     EMP_LIST

------ ------

   10 CLARK     CLARK , KING , MILLER

   10 KING      CLARK , KING , MILLER

   10 MILLER    CLARK , KING , MILLER

   20 SMITH     SMITH , JONES , SCOTT , ADAMS , FORD

   20 JONES     SMITH , JONES , SCOTT , ADAMS , FORD

   20 SCOTT     SMITH , JONES , SCOTT , ADAMS , FORD

   20 ADAMS     SMITH , JONES , SCOTT , ADAMS , FORD

   20 FORD      SMITH , JONES , SCOTT , ADAMS , FORD

   30 ALLEN     ALLEN , WARD , MARTIN , BLAKE , TURNER , JAMES

   30 WARD      ALLEN , WARD , MARTIN , BLAKE , TURNER , JAMES

   30 MARTIN    ALLEN , WARD , MARTIN , BLAKE , TURNER , JAMES

   30 BLAKE     ALLEN , WARD , MARTIN , BLAKE , TURNER , JAMES

   30 TURNER    ALLEN , WARD , MARTIN , BLAKE , TURNER , JAMES

   30 JAMES     ALLEN , WARD , MARTIN , BLAKE , TURNER , JAMES 

14 rows selected