两个分页存储过程的效率比较(均为多表查询)

--解密存储过程
exec sp_decrypt 'p_page'

use Newt8c8
go

if exists(select * from dbo.sysobjects where name='p_page')
drop procedure p_page
go

/**分页的存储过程*/
create procedure p_page
(
@Tables varchar(1000), --表名,可以是多个表,但不能用别名
@PrimaryKey varchar(100), --主键,可以为空,但@Order为空时该值不能为空
@Fields varchar(1000) = '*', --查询字段(多个表的字段),为空表示select *
@PageSize int = 15, --每页记录数
@CurrentPage int = 1, --当前页，表示第页
@Filter varchar(1000) = null,--条件，可以为空，不用填where
@Group varchar(1000) = null, --分组依据，可以为空，不用填group by
@Sort varchar(200) = null, --排序,默认按主键升序排列,不用填order by
@TotalPage int output --总页数
)
with encryption --加密

as
set nocount on
declare @intResult int

begin tran

declare @sql nvarchar(4000)
if @Filter is null or @Filter=''
set @Sql = 'select @intResult = count(' + @PrimaryKey + ') from ' + @Tables
else
set @Sql = 'select @intResult = count(' + @PrimaryKey + ') from ' + @Tables
+ ' where + ' + @Filter

exec sp_executesql @sql,N'@intResult int output',@intResult output
select @TotalPage= ceiling((@intResult+0.0)/@PageSize)

if @Sort is null or @Sort = ''
set @Sort = @PrimaryKey

declare @SortTable varchar(100)
declare @SortName varchar(100)
declare @strSortColumn varchar(200)
declare @operator char(2)
declare @type varchar(100)
declare @prec int

if charindex('desc',@Sort)>0
begin
set @strSortColumn = replace(@Sort, 'desc', '')
set @operator = '<='
end
else
if charindex('asc', @Sort) > 0
begin
set @strSortColumn = replace(@Sort, 'asc', '')
set @operator = '>='
end
else
begin
set @strSortColumn = @SORT
set @operator = '>='
end

if charindex('.', @strSortColumn) > 0
begin
set @SortTable = substring(@strSortColumn, 0, charindex('.',@strSortColumn))
set @SortName=substring(@strSortColumn,charindex('.',@strSortColumn)+1,
len(@strSortColumn))
end
else
begin
set @SortTable = @Tables
set @SortName = @strSortColumn
end

select @type=t.name, @prec=c.prec
from sysobjects o
join syscolumns c on o.id=c.id
join systypes t on c.xusertype=t.xusertype
where o.name = @SortTable and c.name = @SortName

if charindex('char', @type) > 0
set @type = @type + '(' + cast(@prec as varchar) + ')'

declare @strPageSize varchar(50)
declare @strStartRow varchar(50)
declare @strFilter varchar(1000)
declare @strSimpleFilter varchar(1000)
declare @strGroup varchar(1000)

if @CurrentPage < 1
set @CurrentPage = 1

set @strPageSize = cast(@PageSize as varchar(50))
set @strStartRow = cast(((@CurrentPage - 1)*@PageSize + 1) as varchar(50))

if @Filter is not null and @Filter != ''
begin
set @strFilter = ' where ' + @Filter + ' '
set @strSimpleFilter = ' and ' + @Filter + ' '
end
else
begin
set @strSimpleFilter = ''
set @strFilter = ''
end

if @Group is not null and @Group != ''
set @strGroup = ' group by ' + @Group + ' '
else
set @strGroup = ''

set @sql = 'declare @SortColumn ' + @type + ' set rowcount ' + @strStartRow
+ ' select @SortColumn= '+ @strSortColumn + ' from ' + @Tables
+ @strFilter + ' ' + @strGroup + ' order by ' + @Sort + ' set rowcount '
+ @strPageSize + ' select ' + @Fields + ' from ' + @Tables + ' where '
+ @strSortColumn + @operator + ' @SortColumn ' + @strSimpleFilter + ' '
+ @strGroup + ' order by ' + @Sort + ''
exec(@sql)
print @sql
if @@Error <> 0
begin
RollBack Tran
Return -1
end
else
begin
commit Tran
return @intResult
end

select count(1) from song

--测试数据(音乐专辑)
declare
@return_value int,
@TotalPage int,
@timediff datetime
select @timediff=getdate()
exec @return_value=p_page
@Tables='MusicAlbum left outer join Singer on MusicAlbum.SingerID=Singer.RecordID
left outer join MusicArea on Singer.MusicAreaID=MusicArea.RecordID',
@PrimaryKey='MusicAlbum.RecordID',
@Fields='MusicAlbum.RecordID,MusicAlbum.SingerID,MusicAlbum.Name
as AlbumName,Singer.Name as SingerName,MusicArea.Name as MusicArea,
MusicAlbum.SongList,MusicAlbum.IntroductionSmall,MusicAlbum.IssueDate,
MusicAlbum.Label,MusicAlbum.CreateDate',
@PageSize=20,
@CurrentPage=1,
@Filter='',
@Group='',
@Sort='MusicAlbum.CreateDate desc',
@TotalPage=@TotalPage output
select @TotalPage as N'@TotalPage'
select 'Return Value' = @return_value
select datediff(ms,@timediff,GetDate()) as 耗时

--查看CPU用时和占用时间,以此优化存储过程
set statistics io on/off
set statistics time on/off

--测试数据(流行歌曲)
declare
@return_value int,
@TotalPage int,
@timediff datetime
select @timediff=Getdate()
exec @return_value=p_page
@Tables='Song left outer join Singer on Song.SingerIDList=Singer.RecordID
left outer join MusicArea on Singer.MusicAreaID=MusicArea.RecordID',
@PrimaryKey='Song.RecordID',
@Fields='Song.RecordID,MusicArea.Name as Area,Song.Name as songName,
Singer.Name as singerName,Song.State as IsOpen,
Song.CreateDate as Date,Song.EditPeople as Editor',
@PageSize=20,
@CurrentPage=15244,
@Filter='',
@Group='',
@Sort='Song.CreateDate desc',
@TotalPage=@TotalPage output
select @TotalPage as N'@TotalPage'
select 'Return Value' = @return_value
select datediff(ms,@timediff,GetDate()) as 耗时

--查看CPU用时和占用时间,以此优化存储过程
set statistics io on
set statistics time on

set ansi_nulls on
go

set quoted_identifier on
go

use Newt8c8
go

if exists(select * from dbo.sysobjects where name='page_2005')
drop procedure page_2005
go

create proc page_2005
@TableName varchar(500), --表名
@Fields varchar(5000) = '*', --字段名(全部字段为*)
@OrderField varchar(5000), --排序字段(必须!支持多字段)
@sqlWhere varchar(5000) = Null, --条件语句(不用加where)
@pageSize int, --每页多少条记录
@pageIndex int = 1 , --指定当前为第几页
@totalRecord int = 0, --总记录数
@TotalPage int output --返回总页数

as
begin
Begin tran --开始事务
declare @sql nvarchar(4000);
if @totalRecord<=0
begin
--计算总记录数
if (@SqlWhere='' or @sqlWhere=null)
set @sql = 'select @totalRecord = count(*) from ' + @TableName
else
set @sql = 'select @totalRecord = count(*) from '
+ @TableName + ' with(nolock) where ' + @sqlWhere
exec sp_executesql @sql,N'@totalRecord int output',@totalRecord output--计算总记录数
end
--计算总页数
select @TotalPage=ceiling((@totalRecord+0.0)/@PageSize)
if (@SqlWhere='' or @sqlWhere=NULL)
set @sql = 'select * from (select row_number() over(order by ' + @OrderField
+ ') as rowId,' + @Fields + ' from ' + @TableName
else
set @sql = 'select * FROM (select row_number() over(order by ' + @OrderField
+ ') as rowId,' + @Fields + ' from ' + @TableName + ' with(nolock) where '
+ @SqlWhere
--处理页数超出范围情况
if @PageIndex<=0
set @pageIndex = 1
if @pageIndex>@TotalPage
set @pageIndex = @TotalPage
--处理开始点和结束点
declare @StartRecord int
declare @EndRecord int
set @StartRecord = (@pageIndex-1)*@PageSize + 1
set @EndRecord = @StartRecord + @pageSize - 1
--继续合成sql语句
set @Sql = @Sql + ') as t where rowId between ' + Convert(varchar(50),@StartRecord)
+ ' and ' + Convert(varchar(50),@EndRecord)
print @sql
exec(@Sql)
---------------------------------------------------
If @@Error <> 0
begin
rollBack tran
return -1
end
else
begin
commit tran
return @totalRecord ---返回记录总数
end
end
go

--测试数据
declare
@return_value int,
@totalRecord int,
@TotalPage int,
@timediff datetime
select @totalRecord=count(*) from Song
select @timediff=Getdate()
exec @return_value=page_2005
@TableName='Song left outer join Singer on Song.SingerIDList=Singer.RecordID
left outer join MusicArea on Singer.MusicAreaID=MusicArea.RecordID',
@Fields='Song.RecordID,MusicArea.Name as Area,Song.Name as songName,
Singer.Name as singerName,Song.State as IsOpen,
Song.CreateDate as Date,Song.EditPeople as Editor',
@sqlWhere='',
@OrderField='Song.RecordID',
@pageSize=20,
@pageIndex=15244,
@totalRecord=@totalRecord,
@TotalPage=@TotalPage output
select @TotalPage as N'@TotalPage'
select 'Return Value' = @return_value
select datediff(ms,@timediff,GetDate()) as 耗时

p_page和page_2005的分页效率比较(均支持多表查询):
1. 总体查询效率(查询所有数据时):
P_page用时:12898ms
page_2005用时:10256ms
2. 查询前20条数据效率:
P_page用时:1330ms
page_2005用时:3ms
3. 中间查询效率(查询到400条数据时):
P_page用时:1393ms
page_2005用时:76ms
4. 中间查询效率(查询到4000条数据时):
P_page用时:1310ms
page_2005用时:790ms
5. 中间查询效率(查询到40000条数据时):
P_page用时:1266ms
page_2005用时:773ms
6. 中间查询效率(查询到400000条数据时):
P_page用时:890ms
page_2005用时:5936ms
7. 查询到最后一页时:
P_page用时:900ms
page_2005用时:5966ms
总结：如果数据在50万条以上，要求多表（多字段）查询，且最后10页的查询也要快,则使用p_page分页存储过程,反之则用page_2005