选出从m条开始的n条记录

[jronald]

原表中主键ID不连续，现在要选出从m条开始的n条记录，怎样效率最高？
可用SQL Server 2005/2008中的新特性，如ROW_NUMBER等。

[水族杰纶]

1.
select top m * into 临时表(或表变量) from tablename order by columnname -- 将top m笔插入
set rowcount n
select * from 表变量 order by columnname desc


2.
select top n * from
(select top m * from tablename order by columnname) a
order by columnname desc


3.
如果tablename里没有其他identity列，那么：
select identity(int) id0,* into #temp from tablename

取n到m条的语句为：
select * from #temp where id0 >=n and id0 <= m

如果你在执行select identity(int) id0,* into #temp from tablename这条语句的时候报错,那是因为你的DB中间的select into/bulkcopy属性没有打开要先执行：
exec sp_dboption 你的DB名字,'select into/bulkcopy',true


4.
如果表里有identity属性，那么简单：
select * from tablename where identitycol between n and m

[dawugui]

select * from
(
  select * , ROW_NUMBER() OVER(ORDER BY id) px from tb
) t
where px between m and m+n

[dawugui]

取n到m条记录的语句

1.
select top m * from tablename where id not in (select top n id from tablename)

2.
select top m * into 临时表(或表变量) from tablename order by columnname -- 将top m笔插入
set rowcount n
select * from 表变量 order by columnname desc

3.
select top n * from
(select top m * from tablename order by columnname) a
order by columnname desc


4.如果tablename里没有其他identity列，那么：
select identity(int) id0,* into #temp from tablename

取n到m条的语句为：
select * from #temp where id0 >=n and id0 <= m

如果你在执行select identity(int) id0,* into #temp from tablename这条语句的时候报错,那是因为你的DB中间的select into/bulkcopy属性没有打开要先执行：
exec sp_dboption 你的DB名字,'select into/bulkcopy',true


5.如果表里有identity属性，那么简单：
select * from tablename where identitycol between n and m

如果是sql server 2005 可以这样写：
select top 20 * from T order col
except
select top 2 * from T order col

[hyde100]

declare @m smallint
declare @n smallint

set @m = 5
set @n = 10

select * from
(
	select 
		ROW_NUMBER() OVER(ORDER BY item_id) r_id,c_item.* 
	from c_item
) T_test
where r_id between @m and @m + @n