几条复杂的SQL语句

表结构：
CREATE TABLE [dbo].[Exam](
    [S_date] [datetime] NOT NULL,
    [Order_Id] [varchar](50) NOT NULL,
    [Product_Id] [varchar](50) NOT NULL,
    [Amt] [numeric](18, 0) NOT NULL
) ON [PRIMARY]

题目一: 写一条Sql语句查询前出100到199的记录
题目二: 写一条Sql语句删除重复[除时间外的所有字段字段相同]的记录,保留重复记录中时间最大的记录
题目三: 一条Sql语句查出年份,1月,2月,3月....12月的订单总数列表
题目四: 一条sql语句查询出年份,本月销量,上月销量,环比%,去年同期销量,同比%列表

第一题比较简单，前面也讲过，就不赘述。
第二题：
delete from Exam where S_date not in (
    select e2.maxdt from
    (
         select Order_Id,Product_Id,Amt,MAX(S_date) as maxdt from Exam
         group by Order_Id,Product_Id,Amt
    ) as e2
)
第三题：分两次组
select y,sum(c1) as m1,sum(c2) as m2,sum(c3) as m3,sum(c4) as m4,sum(c5) as m5,sum(c6) as m6,
sum(c7) as m7,sum(c8) as m8,sum(c9) as m9,sum(c10) as m10,sum(c11) as m11,sum(c12) as m12
from
(
select
  y,
  case m when 1 then c else 0 end as c1,
  case m when 2 then c else 0 end as c2,
  case m when 3 then c else 0 end as c3,
  case m when 4 then c else 0 end as c4,
  case m when 5 then c else 0 end as c5,
  case m when 6 then c else 0 end as c6,
  case m when 7 then c else 0 end as c7,
  case m when 8 then c else 0 end as c8,
  case m when 9 then c else 0 end as c9,
  case m when 10 then c else 0 end as c10,
  case m when 11 then c else 0 end as c11,
  case m when 12 then c else 0 end as c12
from
  (
   select y,m,count(s_date) as c from
   (
    select datepart(year,convert(DateTime,s_date)) as y,
     datepart(month,convert(DateTime,s_date)) as m ,
     s_date from exam
   )  as T1
   group by T1.y,T1.m
  )
as T2
) as T3
group by T3.y

第三题：SQL相当复杂，不过只要理清了思路就很容易上手了，这里主要用到了left join on，语句如下：

select y1 年,m1 月,c1 本月销售额,
       c2 上月销售额,
       case when c2 is null or c2=0 then '无穷大'
       else cast(cast((isnull(c1, 0)-isnull(c2,0))*100/isnull(c2, 0) as decimal(10,2)) as varchar(50))+'%' end as 环比增长 ,
       c3 去年本月销售额,
       case when c3 is null or c3=0 then '无穷大'
       else cast(cast((isnull(c1, 0)-isnull(c3,0))*100/isnull(c3, 0) as decimal(10,2)) as varchar(50))+'%' end as 同比增长
from
(
select y1,m1,c1,c2,c3
from
(
   select y1,m1,c1,c2 from
   (
    select y1,m1,sum(Amt) as c1 from
     (
          select datepart(year,convert(DateTime,s_date)) as y1,
                   datepart(month,convert(DateTime,s_date)) as m1 ,
                  Amt
         from orders
     )  as T1
     group by T1.y1,T1.m1
    ) o2  

left join
   (
     select y2,m2,sum(Amt) as c2 from
     (
         select datepart(year,convert(DateTime,s_date)) as y2,
                  datepart(month,convert(DateTime,s_date)) as m2 ,
                  Amt from orders
       )  as T1
       group by T1.y2,T1.m2
     ) o3
   on o2.y1 = o3.y2 and o2.m1 = o3.m2 - 1
  ) as o4
  left join
  (
     select y3,m3,sum(Amt) as c3 from
     (
         select datepart(year,convert(DateTime,s_date)) as y3,
                  datepart(month,convert(DateTime,s_date)) as m3,
                  Amt from orders
      )  as T1
      group by T1.y3,T1.m3
   ) as o5
   on o4.y1 = o5.y3 - 1 and o4.m1 = o5.m3
) as o6

 

     年   月    本月销售额           上月销售额   环比        去年同期   同比

    2012 3     230                         420   -45.24%         NULL    无穷大
    2010 4     270                         NULL   无穷大          300      -10.00%
    2011 4     300                         353   -15.01%         420      -28.57%
    2012 4     420                         463   -9.29%           NULL     无穷大
    2011 5     353                         NULL   无穷大          463       -23.76%
    2012 5     463                         NULL   无穷大          NULL     无穷大
    2011 11    492                        240    105.00%        NULL    无穷大
    2011 12    240                        NULL    无穷大          NULL     无穷大