日期:2014-05-18  浏览次数:20682 次

查询问题,在线结分
昨天晚上了个贴子http://topic.csdn.net/u/20071113/20/6abfe1f8-3115-4fed-a2de-0ec449a5c030.html在昨夜小楼和各位朋友的帮助下解决了问题
SQL语句如下
SQL code
 
declare @b table(意见 varchar(20),日期 smalldatetime,状态 int)
insert @b select 'asdf','2007-10-11 0:15:15',0
union all select 'fdasd','2007-10-11 0:12:15' ,1 
union all select 'asasd','2007-10-12 10:15:15'  ,3
union all select 'asds', '2007-10-15 03:12:23'  ,3
union all select 'asfd', '2007-10-17 04:32:12'  ,2

declare @a table(id int identity(0,1),s smalldatetime,e smalldatetime)
declare @s smalldatetime,@e smalldatetime
select @s='2007-10-09',@e='2007-10-20'
insert @a select top 31 null,null from syscolumns

select convert(varchar(10),aa.ss,120) 日期,aa.状态,count(意见) 意见数 from
    (
        select a.*,b.* from
        (select id,dateadd(day,id,@s) SS from @a where dateadd(day,id,@s) <=@e)a
        cross join (select 状态=0 union all select 1 union all select 2 union all select 3) b
    ) aa
left join @b bb
on datediff(day,aa.ss,bb.日期)=0 and aa.状态=bb.状态
group by aa.ss,aa.状态
order by aa.日期 desc,aa.状态


--result
/*
日期          状态          意见数       
---------- -------  -----------
2007-10-09 0 0
2007-10-09 1 0
2007-10-09 2 0
2007-10-09 3 0
2007-10-10 0 0
2007-10-10 1 0
2007-10-10 2 0
2007-10-10 3 0
2007-10-11 0 1
2007-10-11 1 1
2007-10-11 2 0
2007-10-11 3 0
2007-10-12 0 0
2007-10-12 1 0
2007-10-12 2 0
2007-10-12 3 1
2007-10-13 0 0
2007-10-13 1 0
2007-10-13 2 0
2007-10-13 3 0
2007-10-14 0 0
2007-10-14 1 0
2007-10-14 2 0
2007-10-14 3 0
2007-10-15 0 0
2007-10-15 1 0
2007-10-15 2 0
2007-10-15 3 1
2007-10-16 0 0
2007-10-16 1 0
2007-10-16 2 0
2007-10-16 3 0
2007-10-17 0 0
2007-10-17 1 0
2007-10-17 2 1
2007-10-17 3 0
2007-10-18 0 0
2007-10-18 1 0
2007-10-18 2 0
2007-10-18 3 0
2007-10-19 0 0
2007-10-19 1 0
2007-10-19 2 0
2007-10-19 3 0
2007-10-20 0 0
2007-10-20 1 0
2007-10-20 2 0
2007-10-20 3 0

(所影响的行数为 12 行)
*/



现在有一新的要求,想做个统计,就是再结果里面再加两个字段,一个"占比"(该状态的意见数在当天所有状态的意见数中所占的百分比)和一个"比较"(将意见数和前一天的相同状态情况下的意见数作比较(即相减))

------解决方案--------------------
SQL code
declare @b table(意见 varchar(20),日期 smalldatetime,状态 int)
insert @b select 'asdf','2007-10-11 0:15:15',0
union all select 'fdasd','2007-10-11 0:12:15' ,1  
union all select 'asasd','2007-10-12 10:15:15'  ,3 
union all select 'asds', '2007-10-15 03:12:23'   ,3
union all select 'asfd', '2007-10-17 04:32:12'   ,2

declare @a table(id int identity(0,1),s smalldatetime,e smalldatetime)
declare @s smalldatetime,@e smalldatetime
select @s='2007-10-09',@e='2007-10-20'
insert @a select top 31 null,null from syscolumns

select convert(varchar(10),aa.ss,120) 日期,
aa.状态,
count(意见) 意见数 ,
cast(100 * count(意见)/isnull((select sum(1) from @b where datediff(d,日期,aa.ss) = 0 ),1) as varchar)+'%'  占比

from
    (
        select a.*,b.* from
        (select id,dateadd(day,id,@s) SS from @a where dateadd(day,id,@s)<=@e)a
        cross join (select 状态=0 union all select 1 union all select 2 union all select 3) b
    ) aa
left join @b bb
on datediff(day,aa.ss,bb.日期)=0 and aa.状态=bb.状态
group by aa.ss,aa.状态
order by aa.日期 des