Oracle的基于MODEL 子句的行级查询
操作环境服务端Oracle 11g
客户端Oracle 10g
--准备脚本
DROP TABLE EMPLOYEE_SALARY_T;
CREATE TABLE EMPLOYEE_SALARY_T(
EMPLOYEE_ID VARCHAR2(20),
EMPLOYEE_NAME VARCHAR2(20),
DEPARTMENT VARCHAR2(10),
COMPANY VARCHAR2(10),
SALARY NUMBER(6),
SALARY_YEAR VARCHAR2(10)
);
INSERT INTO EMPLOYEE_SALARY_T VALUES('001','znp','dep001','VisSoft',30000,'2005');
INSERT INTO EMPLOYEE_SALARY_T VALUES('001','znp','dep001','VisSoft',40000,'2006');
INSERT INTO EMPLOYEE_SALARY_T VALUES('001','znp','dep001','VisSoft',50000,'2007');
INSERT INTO EMPLOYEE_SALARY_T VALUES('001','znp','dep001','VisSoft',60000,'2008');
INSERT INTO EMPLOYEE_SALARY_T VALUES('002','zsj','dep001','AvsSoft',3000,'2005');
INSERT INTO EMPLOYEE_SALARY_T VALUES('002','zsj','dep001','AvsSoft',4000,'2006');
INSERT INTO EMPLOYEE_SALARY_T VALUES('002','zsj','dep001','AvsSoft',5000,'2007');
INSERT INTO EMPLOYEE_SALARY_T VALUES('002','zsj','dep001','AvsSoft',6000,'2008');
INSERT INTO EMPLOYEE_SALARY_T VALUES('003','ts1','dep002','VisSoft',30000,'2005');
INSERT INTO EMPLOYEE_SALARY_T VALUES('003','ts1','dep002','VisSoft',40000,'2006');
INSERT INTO EMPLOYEE_SALARY_T VALUES('003','ts1','dep002','VisSoft',50000,'2007');
INSERT INTO EMPLOYEE_SALARY_T VALUES('003','ts1','dep002','VisSoft',60000,'2008');
INSERT INTO EMPLOYEE_SALARY_T VALUES('004','ts2','dep002','AvsSoft',3000,'2005');
INSERT INTO EMPLOYEE_SALARY_T VALUES('004','ts2','dep002','AvsSoft',4000,'2006');
INSERT INTO EMPLOYEE_SALARY_T VALUES('004','ts2','dep002','AvsSoft',5000,'2007');
INSERT INTO EMPLOYEE_SALARY_T VALUES('004','ts2','dep002','AvsSoft',6000,'2008');
COMMIT;
--行级查询实现脚本
--查询含义为将09年dep001部门的salary预定为该部门05年、06年与08年的总和
--而dep002部门09年的salary预定为该部门06年、07年与08年的总和
SELECT COMPANY,DEPARTMENT,SALARYS FROM EMPLOYEE_SALARY_T
MODEL
RETURN UPDATED ROWS --是否显示基本查询行,如果有该行则不显示基本查询行，只显示查询计算结果行
PARTITION BY (COMPANY) --分区，对行计算结果的分区，本例中以company列做为分区项
DIMENSION BY (DEPARTMENT,SALARY_YEAR ) --决定两个计算的维度，当前表示第一维度为部门，第二维度为薪资年度
MEASURES (SALARY SALARYS) --我的理解是规则指向列，即对哪一列进行规则计算或规则改变
RULES (
SALARYS['dep001', '2009'] = SALARYS['dep001', '2005'] + SALARYS['dep001', '2006']+ SALARYS['dep001', '2008'],
SALARYS['dep002', '2009'] = SALARYS['dep002', '2006'] + SALARYS['dep002', '2007']+ SALARYS['dep002', '2008']
);
想想如果用列级查询时实现的效果（^_^ 是不是比较地不麻烦）
PS:把COMPANY的值与DEPARTMENT的值调成同步，即DEPARTMENT为001时公司全部为AvsSoft,DEPARTMENT为002时公司全部为VisSoft。看看结果如何？
元旦快乐