Oracle按月份累计求和

原表：

select to_char(reg_Date,'yyyy-mm') regDate, count(*) count from Ep_Info t where 1=1  and t.reg_Date>=to_date('2013-02','yyyy-mm') and t.reg_Date<=to_date('2014-05','yyyy-mm') group by to_char(reg_Date,'yyyy-mm') order by to_char(reg_Date,'yyyy-mm')

按月累计求和：

select to_char(reg_Date,'yyyy-mm') regDate2,sum(count(*)) over(order by to_char(reg_Date,'yyyymm'), count(*)) from Ep_Info t where 1=1    and t.reg_Date>=to_date('2011-12','yyyy-mm') and t.reg_Date<=to_date('2014-04','yyyy-mm') group by to_char(reg_Date,'yyyy-mm'),to_char(reg_Date,'yyyymm') 

参考资料： http://oracle.chinaitlab.com/exploiture/897193.html

--zhongan soft stone

create table zh_an(pnum number(4));

insert into zh_an values(1);

insert into zh_an values(2);

insert into zh_an values(3);

insert into zh_an values(4);

select * from zh_an;

select t.pnum,sum(t.pnum) over(order by t.pnum) leijia from zh_an t;

注意：如果没有order by 子句，求和就不是“连续”的

select t.pnum,sum(t.pnum) over(order by t.pnum) leijia,sum(t.pnum) over() zongshu from zh_an t;

使用的分析函数over (partition by xxx order by xxx1，xxx2 desc) ，在"... from emp;"后面不要加order   by 子句，使用的分析函数的(partition by deptno order by sal)

里已经有排序的语句了，如果再在句尾添加排序子句，一致倒罢了，不一致，结果就令人费劲了。