SQL Server 分页方法汇总

PageSize = 30

PageNumber = 201

方法一:(最常用的分页代码, top
/ not in)

select top 30 UserId from UserInfo where UserId not in (select top 6000 UserId from UserInfo order by UserId) order by UserId

备注: 注意前后的order by
一致

方法二:(not exists, not in
的另一种写法而已)

select top 30 * from UserLog where not exists (select 1 from (select top 6000 LogId from UserLog order by LogId) a where a.LogId = UserLog.LogId) order by LogId

备注:EXISTS用于检查子查询是否至少会返回一行数据,该子查询实际上并不返回任何数据,而是返回值True或False。此处的 select
1 from 也可以是select 2 from,select LogId from, select * from
等等,不影响查询。而且select 1 效率最高,不用查字典表。效率值比较:1
> anycol > *

方法三:(top / max,
局限于使用可比较列排序的时候)

select top 30 * from UserLog where LogId > (select max(LogId) from (select top 6000 LogId from UserLog order by LogId) a ) order by LogId

备注:这里max()函数也可以用于文本列,文本列的比较会根据字母顺序排列,数字
< 字母(无视大小写) < 中文字符

方法四:(row_number() over
(order by LogId))

select top 30 * from ( select row_number() over (order by LogId) as rownumber,* from UserLog)a
where rownumber > 6000 order by LogId

select * from (select row_number()over(order by LogId) as rownumber,* from UserLog)a
where rownumber > 6000 and rownumber < 6030 order by LogId

select * from (select row_number()over(order by LogId) as rownumber,* from UserLog)a
where rownumber between 6000 and  6030 order by LogId

 

select *
from (
    select row_number()over(order by tempColumn)rownumber,*
    from (select top 6030 tempColumn=0,* from UserLog where 1=1 order by LogId)a
)b
where rownumber>6000

row_number() 的变体,不基于已有字段产生记录序号,先按条件筛选以及排好序,再在结果集上给一常量列用于产生记录序号
以上几种方法参考http://www.cnblogs.com/songjianpin/articles/3489050.html

备注:
 
这里rownumber方法属于排名开窗函数(sum, min,
avg等属于聚合开窗函数,ORACLE中叫分析函数,参考文章:SQL SERVER
开窗函数简介 )的一种,搭配over关键字使用。

方法五:(offset /fetch next, SQL
Server 2012支持)

select * from UserLog Order by LogId offset 6000 rows fetch next 30 rows only

备注: 性能参考文章《SQL Server
2012使用OFFSET/FETCH
NEXT分页及性能测试》

 

参考文档:

1、

2、

3、