日期:2014-05-18  浏览次数:20438 次

求SQL语句。如何汇总表中的数据?请诸位高手帮忙!谢谢!
用户个人信息表
userinfo
u_id 用户编号(自动编号)
u_name 用户姓名(如:张三,李四,王五,...)
u_emid 用户合同号(如:123,124,125,...)

如:
u_id u_name u_emid
1 张三 123
2 李四 124
3 王五 125
...


消费表
expense

ep_id 自动编号(1,2,3,...)
ep_u_id 用户编号(取userinfo表u_id字段值)
ep_year 年份(如2010,2011,2012,...)
ep_month 月份(1~12月)
ep_money 月消费

比如现在是2012年4月,expense表数据:
ep_id ep_u_id ep_year ep_month ep_money
71 3 2012 3 1300
70 2 2012 3 1200
69 1 2012 3 1200
.
.
.
37 3 2012 2 800
36 2 2012 2 1100
35 1 2012 2 900
.
.
.
3 3 2012 1 1000
2 2 2012 1 1200
1 1 2012 1 1100


.
.
.

问题:如何汇总输出下列格式?能跨年度算上一年的吗?假如现在是2012.2.1,算出2011.11.1至2012.01.31的汇总?
该汇总数据供查询用。最好用SQL2005及以前版本能用的存储过程来汇总。



汇总ID u_emid u_name 时间段 消费汇总 备注

1 123 张三 2012.01-2012.03 3200.00 2012.01消费1100,2012.2消费900,2012.3消费1200
2 124 李四 2012.01-2012.03 3500.00 2012.01消费1200,2012.2消费1100,2012.3消费1200
3 125 王五 2012.01-2012.03 3100.00 2012.01消费1000,2012.2消费800,2012.3消费1300
...



------解决方案--------------------
SQL code
create proc proc_total
@begindate char(10),
@enddate char(10)
as
--declare @begindate char(10)
--declare @enddate char(10)
--set @begindate='2012.02.01'
--set @enddate='2013.01.01'
select tablea.*,tableb.dept as 备注 from (select distinct u_id, a.u_emid,a.u_name,left(@begindate,7)+'-'+left(@enddate,7) as 时间段,
 sum(ep_money) as 消费汇总
 from userinfo a join expense b on a.u_id=b.ep_u_id where 
left(@begindate,7)<=(ep_year+'.' +(case when len(ep_month)=1 then '0'+ep_month else ep_month end)) 
and left(@enddate,7)>=(ep_year+'.' +(case when len(ep_month)=1 then '0'+ep_month else ep_month end)) 
group by u_id, a.u_emid,a.u_name)tablea join 
(select distinct ep_u_id,dept= stuff((select',' + ep_year+'.'+ep_month+'消费'+cast(ep_money as varchar(20)) from(select ep_u_id,ep_year,ep_month,sum(ep_money) as ep_money from expense where 
left(@begindate,7)<=(ep_year+'.' +(case when len(ep_month)=1 then '0'+ep_month else ep_month end)) 
and left(@enddate,7)>=(ep_year+'.' +(case when len(ep_month)=1 then '0'+ep_month else ep_month end)) group by ep_u_id,ep_year,ep_month
)a where a.ep_u_id=b.ep_u_id  for xml path('')),1,1,'')
from 
(select ep_u_id,ep_year,ep_month,sum(ep_money) as ep_money from expense where 
left(@begindate,7)<=(ep_year+'.' +(case when len(ep_month)=1 then '0'+ep_month else ep_month end)) 
and left(@enddate,7)>=(ep_year+'.' +(case when len(ep_month)=1 then '0'+ep_month else ep_month end)) group by ep_u_id,ep_year,ep_month
)b)tableB on tablea.u_id=tableb.ep_u_id

------解决方案--------------------
SQL code

go
create table userinfo (u_id int,u_name varchar(4),u_emid int)
insert into userinfo
select 1,'张三',123 union all
select 2,'李四',124 union all
select 3,'王五',125

go
create table expense (ep_id int,ep_u_id int,ep_year int,ep_month int,ep_money int)
insert into expense
select 71,3,2012,3,1300 union all
select 70,2,2012,3,1200 union all
select 69,1,2012,3,1200 union all
select 37,3,2012,2,800 union all
select 36,2,2012,2,1100 union all
select 35,1,2012,2,900 union all
select 3,3,2012,1,1000 union all
select 2,2,2012,1,1200 union all
select 1,1,2012,1,1100

go

create proc proc_expense
(
    @username varchar(20),
    @begintime datetime,
    @endtime datetime
)
as
begin
    ;with maco as(
    select 
        ep_u_id as u_emid,
        (select top 1 u_name from userinfo where u_id=t.ep_u_id) as u_name,
         replace(convert(varchar(7),@begintime,111)+'-'+convert(varchar(7),@endtime,111),'/','.') as 时间段,
        (select sum(ep_money) from expense where ep_u_id=t.ep_u_id) as 消费汇总,