数据统计的问题----------特请邹建大哥指教

[jinhai2003]

现在我要将今天早上08:00:00到明天早上07:59:59之间的所有数据（字段DTRN）求和汇总后生成一报表，汇总后的结果表示今天的累计值
表名：A
表结构及表记录如下：
YMDHM DTRN
2004-04-01 07:59:59.000 .0
2004-04-01 09:40:06.000 .0
2004-04-01 19:59:59.000 .3
2004-04-01 22:40:06.000 .5
2004-04-02 05:50:42.000 .5
2004-04-02 06:45:48.000 .5
2004-04-02 07:45:42.000 .5
2004-04-02 07:59:59.000 .4
2004-04-02 08:12:22.000 .5
2004-04-02 08:42:23.000 .5
2004-04-02 09:02:31.000 .5
2004-04-02 09:37:39.000 .5
2004-04-02 11:17:18.000 1.5
2004-04-02 13:09:27.000 .5
2004-04-02 15:40:06.000 .0
2004-04-02 19:59:59.000 .3
2004-04-02 22:40:06.000 .5
2004-04-02 23:53:06.000 .5
2004-04-03 04:43:00.000 2.5
2004-04-03 07:59:59.000 1.2
2004-04-03 09:59:37.000 .5
2004-04-03 19:59:59.000 .2
2004-04-03 22:40:06.000 .5
2004-04-04 03:52:43.000 .5
2004-04-04 04:47:34.000 .5
2004-04-04 05:26:58.000 .5
2004-04-04 06:39:36.000 .5
2004-04-04 07:59:59.000 .0
统计后的报表应为：
YMDHM DTRN
2004-04-01 2.7
2004-04-02 9.0
2004-04-03 3.2
请问用存储过程能否实现，或用其他方法能实现请各位大哥不吝赐教，多谢了！！

[zjcxc]

select convert(varchar(10),dateadd(hour,8,YMDHM),120),DTRN=sum(DTRN)
from 表
group by DTRN

[zjcxc]

--上面写错了
select YMDHM=convert(varchar(10),dateadd(hour,-8,YMDHM),120),DTRN=sum(DTRN)
from a
group by convert(varchar(10),dateadd(hour,-8,YMDHM),120)

[zjcxc]

--测试

--测试数据
create table a(YMDHM datetime,DTRN decimal(5,1))
insert a select '2004-04-01 07:59:59.000',.0
union all select '2004-04-01 09:40:06.000',.0
union all select '2004-04-01 19:59:59.000',.3
union all select '2004-04-01 22:40:06.000',.5
union all select '2004-04-02 05:50:42.000',.5
union all select '2004-04-02 06:45:48.000',.5
union all select '2004-04-02 07:45:42.000',.5
union all select '2004-04-02 07:59:59.000',.4
union all select '2004-04-02 08:12:22.000',.5
union all select '2004-04-02 08:42:23.000',.5
union all select '2004-04-02 09:02:31.000',.5
union all select '2004-04-02 09:37:39.000',.5
union all select '2004-04-02 11:17:18.000',1.5
union all select '2004-04-02 13:09:27.000',.5
union all select '2004-04-02 15:40:06.000',.0
union all select '2004-04-02 19:59:59.000',.3
union all select '2004-04-02 22:40:06.000',.5
union all select '2004-04-02 23:53:06.000',.5
union all select '2004-04-03 04:43:00.000',2.5
union all select '2004-04-03 07:59:59.000',1.2
union all select '2004-04-03 09:59:37.000',.5
union all select '2004-04-03 19:59:59.000',.2
union all select '2004-04-03 22:40:06.000',.5
union all select '2004-04-04 03:52:43.000',.5
union all select '2004-04-04 04:47:34.000',.5
union all select '2004-04-04 05:26:58.000',.5
union all select '2004-04-04 06:39:36.000',.5
union all select '2004-04-04 07:59:59.000',.0
go

--统计
select YMDHM=convert(varchar(10),dateadd(hour,-8,YMDHM),120),DTRN=sum(DTRN)
from a
group by convert(varchar(10),dateadd(hour,-8,YMDHM),120)
go

--删除测试
drop table a

/*--测试结果
YMDHM DTRN
---------- ----------------------------------------
2004-03-31 .0
2004-04-01 2.7
2004-04-02 9.0
2004-04-03 3.2

（所影响的行数为 4 行）
--*/

[arrow_gx]

select DATEPART (yyyy, YMDHM ) ,DATEPART (mm, YMDHM ),DATEPART (dd, YMDHM ),sum(dtrm) from table group by DATEPART (yyyy, YMDHM ) ,DATEPART (mm, YMDHM ),DATEPART (dd, YMDHM )

[outwindows]

DTRN 为 float 型时为什么结果是这样的呢？
YMDHM DTRN
---------- ----------------------------------------
2004-03-31 0.0
2004-04-01 2.6999999999999997
2004-04-02 9.0
2004-04-03 3.2000000000000002

[zjcxc]

float是非精度类型,结果当然会这样.