有没有高效的分页存储过程
CREATE   procedure   PageChange
(
@pagesize   int,
@pageindex   int,
@Keywords   varchar(50),
@docount   bit
)
as
set   nocount   on
if(@docount=1)
select   count(*)   from   table1   as   t1   inner   join   table2   on   t1.uID=table2.uID   inner   join   containstable(table3,Content,@Keywords)   as   t3   on   t1.ID=t3.[KEY]
else
begin
declare   @indextable   table(id   int   identity(1,1),nid   int)
declare   @PageLowerBound   int
declare   @PageUpperBound   int
set   @PageLowerBound=(@pageindex-1)*@pagesize
set   @PageUpperBound=@PageLowerBound+@pagesize
set   rowcount   @PageUpperBound

insert   into   @indextable(nid)   select   ID   from   table1   as   t1   inner   join   table2   on   t1.uID=table2.uID   inner   join   containstable(table1,Content,@Keywords)   as   t3
on   t1.ID=t3.[KEY]   order   by   t3.[RANK]   desc,t1.AddTime   desc

insert   into   @indextable(nid)....
......
......

    SELECT   ...
                  ...
    from   table1   O,@indextable   t   ,table2
where   O.ID=t.nid
and   t.id> @PageLowerBound   and   t.id <=@PageUpperBound     and   O.uID=table2.uID   order   by   t.id

end
set   nocount   off
GO


这个存储过程是可以实现功能，但是数据量非常大的时候，查询速度就比较慢了。
不知有哪位高手能提供更好的分页存储过程？

------解决方案--------------------
http://topic.csdn.net/t/20031017/11/2365596.html
------解决方案--------------------
select ID from table1 as t1 inner join table2 on t1.uID=table2.uID inner join containstable(table1,Content,@Keywords) as t3
on t1.ID=t3.[KEY] order by t3.[RANK] desc,t1.AddTime desc

修改一下，table2有意义么？就查询ID来说应该无意义，去掉。