求sql，先分组后批改值

求sql，先分组后修改值
求sql，先分组后批改值
以id分组 date 排序，分组后每组第一条的 no 不变，从第二条开始原始no - 前一行的no = 现在的no
结果大概是：
1002013-01-29 9
10062013-02-17 9
10062013-02-28 7
10062013-03-12 20

请问这个sql应该怎么写？
[解决办法]


select IDENTITY(int,1,1)pid,* into #t from(
select 100 id,convert(date,'2013-01-29')date,  9 no
union all select 1006, '2013-02-17',  9
union all select 1006, '2013-02-28',  16
union all select 1006, '2013-03-12',  36
)a

select a.*,a.no-isnull(b.no,0) from #t a left join #t b on a.id=b.id and a.pid=b.pid+1

/*
pididdateno(无列名)
11002013-01-2999
210062013-02-1799
310062013-02-28167
410062013-03-123620
*/

[解决办法]

select a.id,convert(nvarchar(10),a.date,120) date ,a.no - isnull(b.no,0) no
from  (
select id,row_number() over (partition by id order by pid ) num,date,no
from test_demo) a left join (
select id,row_number() over (partition by id order by pid ) num,date,no
from test_demo)
b on a.id = b.id and a.num = b.num + 1 
/*
id,date,no
100,2013-01-29,9
100,2013-03-28,19
1006,2013-02-17,9
1006,2013-02-28,7
1006,2013-03-12,20

(5 行受影响)

[解决办法]
楼主的问题是“先分组后修改值”，请问还要修改原始表的记录吗？
如果要的话，则：
select IDENTITY(int,1,1)pid,id,[date],[no] into #t from Test_Demo
update Test_Demo t1
set t1.[no]=t2.No_upd
from (select a.*,a.no-isnull(b.no,0) as No_upd from #t a left join #t b on a.id=b.id and a.pid=b.pid+1)t2
where t1.id=t2.id and t1.[date]=t2.[date]

--以后建表时，不要使用SQL server的保留字了（如date等）
[解决办法]

--数据临测表
CREATE TABLE Test_Demo
(
pid int primary key identity(1,1),--主键，自增1
id  nvarchar(20) not null,--组
[date] datetime not null,--日期
[no]int not null--编号
)
GO

INSERT INTO  Test_Demo SELECT '100','2013-01-29','9'
UNION ALL SELECT '1006','2013-02-17','9' 
UNION ALL SELECT '1006','2013-02-28','16' 
UNION ALL SELECT '1006','2013-03-12','36' 
UNION ALL SELECT '100','2013-03-28','28' 

select m.id , m.date , m.no - isnull((select top 1 no from test_demo n where n.id = m.id and n.pid &lt; m.pid order by n.pid desc),0) from test_demo m
order by m.id , m.pid

drop table test_demo

/*
id                   date                                                                
 
-------------------- ------------------------------------------------------ ----------- 
100                  2013-01-29 00:00:00.000                                9
100                  2013-03-28 00:00:00.000                                19
1006                 2013-02-17 00:00:00.000                                9
1006                 2013-02-28 00:00:00.000                                7
1006                 2013-03-12 00:00:00.000                                20

（所影响的行数为 5 行）
*/

求sql先分组后批改值

热点推荐