Oracle中Keep的使用

官方文档有如下说明：

FIRST/LAST Functions

The FIRST/LAST aggregate functions allow you to return the result of an aggregate applied over a set of rows that rank as the first or last with respect to a given order specification. FIRST/LAST lets you order on column A but return an result of an aggregate applied on column B. This is valuable because it avoids the need for a self-join or subquery, thus improving performance. These functions begin with a tiebreaker function, which is a regular aggregate function (MIN, MAX, SUM, AVG, COUNT, VARIANCE, STDDEV) that produces the return value. The tiebreaker function is performed on the set rows (1 or more rows) that rank as first or last respect to the order specification to return a single value.

To specify the ordering used within each group, the FIRST/LAST functions add a new clause starting with the word KEEP.

FIRST/LAST Syntax

These functions have the following syntax:

aggregate_function KEEP
( DENSE_RANK LAST ORDER BY
  expr [ DESC | ASC ] [NULLS { FIRST | LAST }]
  [, expr [ DESC | ASC ] [NULLS { FIRST | LAST }]]...
)
[OVER query_partitioning_clause]

-- emp表的数据
SQL> SELECT t.empno,
  2         t.ename,
  3         t.mgr,
  4         t.sal,
  5         t.deptno
  6    FROM emp t
  7   ORDER BY t.sal,
  8            t.deptno;

     EMPNO ENAME                       MGR        SAL     DEPTNO
---------- -------------------- ---------- ---------- ----------
       111 aaa                        2222        800          9
      7369 SMITH                      7902        800         20
      7900 JAMES                      7698        950         30
      7876 ADAMS                      7788       1100         20
      7521 WARD                       7698       1250         30
      7654 MARTIN                     7698       1250         30
      7934 MILLER                     7782       1300         10
      7844 TURNER                     7698       1500         30
      7499 ALLEN                      7698       1600         30
      7782 CLARK                      7839       2450         10
      7698 BLAKE                      7839       2850         30

     EMPNO ENAME                       MGR        SAL     DEPTNO
---------- -------------------- ---------- ---------- ----------
      7566 JONES                      7839       2975         20
      7788 SCOTT                      7566       3000         20
      7902 FORD                       7566       3000         20
      7839 KING                                  5000         10
       222 bbb                        3333       5000         40

-- 1.现在要查询表中工资最高的部门号的最大最小值,工资最低的部门号的最大最小值
-- 因为是DENSE_RANK，会产生重复数据，使用min,max取一条。
-- 这个sql没有使用over子句，后面的例子会使用
SQL> SELECT MIN(t.deptno) KEEP(DENSE_RANK FIRST ORDER BY t.sal) a,
  2         MAX(t.deptno) KEEP(DENSE_RANK FIRST ORDER BY t.sal) b,
  3         MIN(t.deptno) KEEP(DENSE_RANK LAST ORDER BY t.sal) c,
  4         MAX(t.deptno) KEEP(DENSE_RANK LAST ORDER BY t.sal) d
  5    FROM emp t;

         A          B          C          D
---------- ---------- ---------- ----------
         9         20         10         40

-- 2.加上over,对每一行记录做计算，看看效果：
SQL> 
SQL> SELECT t.empno,
  2         t.ename,
  3         t.mgr,
  4         t.sal,
  5         t.deptno,
  6         MIN(t.deptno) KEEP(DENSE_RANK FIRST ORDER BY t.sal) OVER() a,
  7         MAX(t.deptno) KEEP(DENSE_RANK FIRST ORDER BY t.sal) OVER() b,
  8         MIN(t.deptno) KEEP(DENSE_RANK LAST ORDER BY t.sal) OVER() c,
  9         MAX(t.deptno) KEEP(DENSE_RANK LAST ORDER BY t.sal) OVER() d
 10    FROM emp t
 11   ORDER BY t.sal,
 12            t.deptno
 13  ;

EMPNO ENAME        MGR       SAL DEPTNO          A          B          C          D
----- ---------- ----- --------- ------ ---------- ---------- ---------- ----------
  111 aaa         2222    800.00      9          9         20         10         40
 7369 SMITH       7902    800.00     20          9         20         10         40
 7900 JAMES       7698    950.00     30          9         20         10         40
 7876 ADAMS       7788   1100.00     20          9         20         10         40
 7521 WARD        7698   1250.00     30          9         20         10         40
 7654 MARTIN      7698   1250.00     30          9         20         10         40
 7934 MILLER      7782   1300.00     10          9         20         10         40
 7844 TURNER      7698   1500.00     30          9         20         10         40
 7499 ALLEN       7698   1600.00     30          9         20         10         40
 7782 CLARK       7839   2450.00     10          9         20         10         40
 7698 BLAKE       7839   2850.00     30          9         20         10         40
 7566 JONES       7839   2975.00     20          9         20         10         40
 7788 SCOTT       7566   3000.00     20          9         20         10         40
 7902 FORD        7566   3000.00     20          9         20         10         40
 7839 KING               5000.00     10          9         20         10         40
  222 bbb         3333   5000.00     40          9         20         10         40

-- 3.下面对每一个mgr求最大(最小)工资的部门号的最大(最小)值
SQL> SELECT t.empno,
  2         t.ename,
  3         t.mgr,
  4         t.sal,
  5         t.deptno,
  6         MIN(t.deptno) KEEP(DENSE_RANK FIRST ORDER BY t.sal) OVER(PARTITION BY t.mgr) a,
  7         MAX(t.deptno) KEEP(DENSE_RANK FIRST ORDER BY t.sal) OVER(PARTITION BY t.mgr) b,
  8         MIN(t.deptno) KEEP(DENSE_RANK LAST ORDER BY t.sal) OVER(PARTITION BY t.mgr) c,
  9         MAX(t.deptno) KEEP(DENSE_RANK LAST ORDER BY t.sal) OVER(PARTITION BY t.mgr) d
 10    FROM emp t
 11   ORDER BY t.sal,
 12            t.deptno
 13  ;

EMPNO ENAME        MGR       SAL DEPTNO          A          B          C          D
----- ---------- ----- --------- ------ ---------- ---------- ---------- ----------
  111 aaa         2222    800.00      9          9          9          9          9
 7369 SMITH       7902    800.00     20         20         20         20         20
 7900 JAMES       7698    950.00     30         30         30         30         30
 7876 ADAMS       7788   1100.00     20         20         20         20         20
 7654 MARTIN      7698   1250.00     30         30         30         30         30
 7521 WARD        7698   1250.00     30         30         30         30         30
 7934 MILLER      7782   1300.00     10         10         10         10         10
 7844 TURNER      7698   1500.00     30         30         30         30         30
 7499 ALLEN       7698   1600.00     30         30         30         30         30
 7782 CLARK       7839   2450.00     10         10         10         20         20
 7698 BLAKE       7839   2850.00     30         10         10         20         20
 7566 JONES       7839   2975.00     20         10         10         20         20
 7902 FORD        7566   3000.00     20         20         20         20         20
 7788 SCOTT       7566   3000.00     20         20         20         20         20
 7839 KING               5000.00     10         10         10         10         10
  222 bbb         3333   5000.00     40         40         40         40         40