在项目中，我们经常遇到或用到分页，那么在大数据量（百万级以上）下，哪种分页算法效率最优呢？我们不妨用事实说话。
测试环境
硬件：CPU 酷睿双核T5750 内存：2G
软件:Windows server 2003 + Sql server 2005
OK,我们首先创建一数据库：data_Test,并在此数据库中创建一表：tb_TestTable
1 create database data_Test -- 创建数据库data_Test 2 GO 3 use data_Test 4 GO 5 create table tb_TestTable -- 创建表 6 ( 7 id int identity ( 1 , 1 ) primary key , 8 userName nvarchar ( 20 ) not null , 9 userPWD nvarchar ( 20 ) not null , 10 userEmail nvarchar ( 40 ) null 11 ) 12 GO
然后我们在数据表中插入2000000条数据:
1 -- 插入数据 2 set identity_insert tb_TestTable on 3 declare @count int 4 set @count = 1 5 while @count <= 2000000 6 begin 7 insert into tb_TestTable(id,userName,userPWD,userEmail) values ( @count , ' admin ' , ' admin888 ' , ' lli0077@yahoo.com.cn ' ) 8 set @count = @count + 1 9 end 10 set identity_insert tb_TestTable off
我首先写了五个常用存储过程：
1，利用select top 和select not in进行分页，具体代码如下：
1 create procedure proc_paged_with_notin -- 利用select top and select not in 2 ( 3 @pageIndex int , -- 页索引 4 @pageSize int -- 每页记录数 5 ) 6 as 7 begin 8 set nocount on ; 9 declare @timediff datetime -- 耗时 10 declare @sql nvarchar ( 500 ) 11 select @timediff = Getdate () 12 set @sql = ' select top ' + str ( @pageSize ) + ' * from tb_TestTable where(ID not in(select top ' + str ( @pageSize * @pageIndex ) + ' id from tb_TestTable order by ID ASC)) order by ID ' 13 execute ( @sql ) -- 因select top后不支技直接接参数，所以写成了字符串@sql 14 select datediff (ms, @timediff , GetDate ()) as 耗时 15 set nocount off ; 16 end
2,利用select top 和 select max(列键)
1 create procedure proc_paged_with_selectMax -- 利用select top and select max(列) 2 ( 3 @pageIndex int , -- 页索引 4 @pageSize int -- 页记录数 5 ) 6 as 7 begin 8 set nocount on ; 9 declare @timediff datetime 10 declare @sql nvarchar ( 500 ) 11 select @timediff = Getdate () 12 set @sql = ' select top ' + str ( @pageSize ) + ' * From tb_TestTable where(ID>(select max(id) From (select top ' + str ( @pageSize * @pageIndex ) + ' id From tb_TestTable order by ID) as TempTable)) order by ID ' 13 execute ( @sql ) 14 select datediff (ms, @timediff , GetDate ()) as 耗时 15 set nocount off ; 16 end
3,利用select top和中间变量--此方法因网上有人说效果最佳，所以贴出来一同测试
1 create procedure proc_paged_with_Midvar -- 利用ID>最大ID值和中间变量 2 ( 3 @pageIndex int , 4 @pageSize int 5 ) 6 as 7 declare @count int 8 declare @ID int 9 declare @timediff datetime 10 declare @sql nvarchar ( 500 ) 11 begin 12 set nocount on ; 13 select @count = 0 , @ID = 0 , @timediff = getdate () 14 select @count = @count + 1 , @ID = case when @count <= @pageSize * @pageIndex then ID else @ID end from tb_testTable order by id 15 set @sql = ' select top ' + str ( @pageSize ) + ' * from tb_testTable where ID> ' + str ( @ID ) 16 execute ( @sql ) 17 select datediff (ms, @timediff , getdate ()) as 耗时 18 set nocount off ; 19 end 20
4,利用Row_number() 此方法为SQL server 2005中新的方法,利用Row_number()给数据行加上索引
1 create procedure proc_paged_with_Rownumber -- 利用SQL 2005中的Row_number() 2 ( 3 @pageIndex int , 4 @pageSize int 5 ) 6 as 7 declare @timediff datetime 8 begin 9 set nocount on ; 10 select @timediff = getdate () 11 select * from ( select * ,Row_number() over ( order by ID asc ) as IDRank from tb_testTable) as IDWithRowNumber where IDRank > @pageSize * @pageIndex and IDRank < @pageSize * ( @pageIndex + 1 ) 12 select datediff (ms, @timediff , getdate ()) as 耗时 13 set nocount off ; 14 end 15
5，利用临时表及Row_number
1 create procedure proc_CTE -- 利用临时表及Row_number 2 ( 3 @pageIndex int , -- 页索引 4 @pageSize int -- 页记录数 5 ) 6 as 7 set nocount on ; 8 declare @ctestr nvarchar ( 400 ) 9 declare @strSql nvarchar ( 400 ) 10 declare @datediff datetime 11 begin 12 select @datediff = GetDate () 13 set @ctestr = ' with Table_CTE as 14 (select ceiling((Row_number() over(order by ID ASC))/ ' + str ( @pageSize ) + ' ) as page_num,* from tb_TestTable) ' ; 15 set @strSql = @ctestr + ' select * From Table_CTE where page_num= ' + str ( @pageIndex ) 16 end 17 begin 18 execute sp_executesql @strSql 19 select datediff (ms, @datediff , GetDate ()) 20 set nocount off ; 21 end 22
OK,至此，存储过程创建完毕，我们分别在每页10条数据的情况下在第2页，第1000页，第10000页，第100000页，第199999页进行测试，耗时单位：ms 每页测试5次取其平均值
存过第2页耗时第1000页耗时第10000页耗时第100000页耗时第199999页耗时效率排行1用not in0ms16ms47ms475ms953ms32用select max5ms16ms35ms325ms623ms13中间变量966ms970ms960ms945ms933ms54row_number0ms0ms34ms365ms710ms24临时表780ms796ms798ms780ms805ms4
测试结果显示：select max >row_number>not in>临时表>中间变量
于是我对效率最高的select max方法用2分法进行了扩展，代码取自互联网，我修改了ASC排序时取不到值的BUG，测试结果:
2分法156ms156ms180ms470ms156ms1*
从测试结果来看，使用2分法确实可以提高效率并使效率更为稳定，我又增加了第159999页的测试，用时仅296ms，效果相当的不错！
下面是2分法使用select max的代码，已相当完善。
1 -- /*-----存储过程 分页处理 孙伟 2005-03-28创建 -------*/ 2 -- /*-----存储过程 分页处理 浪尘 2008-9-1修改----------*/ 3 -- /*----- 对数据进行了2分处理使查询前半部分数据与查询后半部分数据性能相同 -------*/ 4 5 alter PROCEDURE proc_paged_2part_selectMax 6 ( 7 @tblName nvarchar ( 200 ), -- --要显示的表或多个表的连接 8 @fldName nvarchar ( 500 ) = ' * ' , -- --要显示的字段列表 9 @pageSize int = 10 , -- --每页显示的记录个数 10 @page int = 1 , -- --要显示那一页的记录 11 @fldSort nvarchar ( 200 ) = null , -- --排序字段列表或条件 12 @Sort bit = 0 , -- --排序方法，0为升序，1为降序(如果是多字段排列Sort指代最后一个排序字段的排列顺序(最后一个排序字段不加排序标记)--程序传参如：' SortA Asc,SortB Desc,SortC ') 13 @strCondition nvarchar ( 1000 ) = null , -- --查询条件,不需where 14 @ID nvarchar ( 150 ), -- --主表的主键 15 @Dist bit = 0 , -- --是否添加查询字段的 DISTINCT 默认0不添加/1添加 16 @pageCount int = 1 output, -- --查询结果分页后的总页数 17 @Counts int = 1 output -- --查询到的记录数 18 ) 19 AS 20 SET NOCOUNT ON 21 Declare @sqlTmp nvarchar ( 1000 ) -- --存放动态生成的SQL语句 22 Declare @strTmp nvarchar ( 1000 ) -- --存放取得查询结果总数的查询语句 23 Declare @strID nvarchar ( 1000 ) -- --存放取得查询开头或结尾ID的查询语句 24 25 Declare @strSortType nvarchar ( 10 ) -- --数据排序规则A 26 Declare @strFSortType nvarchar ( 10 ) -- --数据排序规则B 27 28 Declare @SqlSelect nvarchar ( 50 ) -- --对含有DISTINCT的查询进行SQL构造 29 Declare @SqlCounts nvarchar ( 50 ) -- --对含有DISTINCT的总数查询进行SQL构造 30 31 declare @timediff datetime -- 耗时测试时间差 32 select @timediff = getdate () 33 34 if @Dist = 0 35 begin 36 set @SqlSelect = ' select ' 37 set @SqlCounts = ' Count(*) ' 38 end 39 else 40 begin 41 set @SqlSelect = ' select distinct ' 42 set @SqlCounts = ' Count(DISTINCT ' + @ID + ' ) ' 43 end 44 45 46 if @Sort = 0 47 begin 48 set @strFSortType = ' ASC ' 49 set @strSortType = ' DESC ' 50 end 51 else 52 begin 53 set @strFSortType = ' DESC ' 54 set @strSortType = ' ASC ' 55 end 56 57 58 59 -- ------生成查询语句-------- 60 -- 此处@strTmp为取得查询结果数量的语句 61 if @strCondition is null or @strCondition = '' -- 没有设置显示条件 62 begin 63 set @sqlTmp = @fldName + ' From ' + @tblName 64 set @strTmp = @SqlSelect + ' @Counts= ' + @SqlCounts + ' FROM ' + @tblName 65 set @strID = ' From ' + @tblName 66 end 67 else 68 begin 69 set @sqlTmp = + @fldName + ' From ' + @tblName + ' where (1>0) ' + @strCondition 70 set @strTmp = @SqlSelect + ' @Counts= ' + @SqlCounts + ' FROM ' + @tblName + ' where (1>0) ' + @strCondition 71 set @strID = ' From ' + @tblName + ' where (1>0) ' + @strCondition 72 end 73 74 -- --取得查询结果总数量----- 75 exec sp_executesql @strTmp ,N ' @Counts int out ' , @Counts out 76 declare @tmpCounts int 77 if @Counts = 0 78 set @tmpCounts = 1 79 else 80 set @tmpCounts = @Counts 81 82 -- 取得分页总数 83 set @pageCount = ( @tmpCounts + @pageSize - 1 ) / @pageSize 84 85 /**/ /**/ /**/ /**当前页大于总页数 取最后一页**/ 86 if @page > @pageCount 87 set @page = @pageCount 88 89 -- /*-----数据分页2分处理-------*/ 90 declare @pageIndex int -- 总数/页大小 91 declare @lastcount int -- 总数%页大小 92 93 set @pageIndex = @tmpCounts / @pageSize 94 set @lastcount = @tmpCounts % @pageSize 95 if @lastcount > 0 96 set @pageIndex = @pageIndex + 1 97 else 98 set @lastcount = @pagesize 99 100 -- //***显示分页 101 if @strCondition is null or @strCondition = '' -- 没有设置显示条件 102 begin 103 if @pageIndex < 2 or @page <= @pageIndex / 2 + @pageIndex % 2 -- 前半部分数据处理 104 begin 105 if @page = 1 106 set @strTmp = @SqlSelect + ' top ' + CAST ( @pageSize as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 107 + ' order by ' + @fldSort + ' ' + @strFSortType 108 else 109 begin 110 if @Sort = 1 111 begin 112 set @strTmp = @SqlSelect + ' top ' + CAST ( @pageSize as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 113 + ' where ' + @ID + ' <(select min( ' + @ID + ' ) from ( ' + @SqlSelect + ' top ' + CAST ( @pageSize * ( @page - 1 ) as Varchar ( 20 )) + ' ' + @ID + ' from ' + @tblName 114 + ' order by ' + @fldSort + ' ' + @strFSortType + ' ) AS TBMinID) ' 115 + ' order by ' + @fldSort + ' ' + @strFSortType 116 end 117 else 118 begin 119 set @strTmp = @SqlSelect + ' top ' + CAST ( @pageSize as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 120 + ' where ' + @ID + ' >(select max( ' + @ID + ' ) from ( ' + @SqlSelect + ' top ' + CAST ( @pageSize * ( @page - 1 ) as Varchar ( 20 )) + ' ' + @ID + ' from ' + @tblName 121 + ' order by ' + @fldSort + ' ' + @strFSortType + ' ) AS TBMinID) ' 122 + ' order by ' + @fldSort + ' ' + @strFSortType 123 end 124 end 125 end 126 else 127 begin 128 set @page = @pageIndex - @page + 1 -- 后半部分数据处理 129 if @page <= 1 -- 最后一页数据显示 130 set @strTmp = @SqlSelect + ' * from ( ' + @SqlSelect + ' top ' + CAST ( @lastcount as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 131 + ' order by ' + @fldSort + ' ' + @strSortType + ' ) AS TempTB ' + ' order by ' + @fldSort + ' ' + @strFSortType 132 else 133 if @Sort = 1 134 begin 135 set @strTmp = @SqlSelect + ' * from ( ' + @SqlSelect + ' top ' + CAST ( @pageSize as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 136 + ' where ' + @ID + ' >(select max( ' + @ID + ' ) from( ' + @SqlSelect + ' top ' + CAST ( @pageSize * ( @page - 2 ) + @lastcount as Varchar ( 20 )) + ' ' + @ID + ' from ' + @tblName 137 + ' order by ' + @fldSort + ' ' + @strSortType + ' ) AS TBMaxID) ' 138 + ' order by ' + @fldSort + ' ' + @strSortType + ' ) AS TempTB ' + ' order by ' + @fldSort + ' ' + @strFSortType 139 end 140 else 141 begin 142 set @strTmp = @SqlSelect + ' * from ( ' + @SqlSelect + ' top ' + CAST ( @pageSize as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 143 + ' where ' + @ID + ' <(select min( ' + @ID + ' ) from( ' + @SqlSelect + ' top ' + CAST ( @pageSize * ( @page - 2 ) + @lastcount as Varchar ( 20 )) + ' ' + @ID + ' from ' + @tblName 144 + ' order by ' + @fldSort + ' ' + @strSortType + ' ) AS TBMaxID) ' 145 + ' order by ' + @fldSort + ' ' + @strSortType + ' ) AS TempTB ' + ' order by ' + @fldSort + ' ' + @strFSortType 146 end 147 end 148 end 149 150 else -- 有查询条件 151 begin 152 if @pageIndex < 2 or @page <= @pageIndex / 2 + @pageIndex % 2 -- 前半部分数据处理 153 begin 154 if @page = 1 155 set @strTmp = @SqlSelect + ' top ' + CAST ( @pageSize as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 156 + ' where 1=1 ' + @strCondition + ' order by ' + @fldSort + ' ' + @strFSortType 157 else if ( @Sort = 1 ) 158 begin 159 set @strTmp = @SqlSelect + ' top ' + CAST ( @pageSize as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 160 + ' where ' + @ID + ' <(select min( ' + @ID + ' ) from ( ' + @SqlSelect + ' top ' + CAST ( @pageSize * ( @page - 1 ) as Varchar ( 20 )) + ' ' + @ID + ' from ' + @tblName 161 + ' where (1=1) ' + @strCondition + ' order by ' + @fldSort + ' ' + @strFSortType + ' ) AS TBMinID) ' 162 + ' ' + @strCondition + ' order by ' + @fldSort + ' ' + @strFSortType 163 end 164 else 165 begin 166 set @strTmp = @SqlSelect + ' top ' + CAST ( @pageSize as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 167 + ' where ' + @ID + ' >(select max( ' + @ID + ' ) from ( ' + @SqlSelect + ' top ' + CAST ( @pageSize * ( @page - 1 ) as Varchar ( 20 )) + ' ' + @ID + ' from ' + @tblName 168 + ' where (1=1) ' + @strCondition + ' order by ' + @fldSort + ' ' + @strFSortType + ' ) AS TBMinID) ' 169 + ' ' + @strCondition + ' order by ' + @fldSort + ' ' + @strFSortType 170 end 171 end 172 else 173 begin 174 set @page = @pageIndex - @page + 1 -- 后半部分数据处理 175 if @page <= 1 -- 最后一页数据显示 176 set @strTmp = @SqlSelect + ' * from ( ' + @SqlSelect + ' top ' + CAST ( @lastcount as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 177 + ' where (1=1) ' + @strCondition + ' order by ' + @fldSort + ' ' + @strSortType + ' ) AS TempTB ' + ' order by ' + @fldSort + ' ' + @strFSortType 178 else if ( @Sort = 1 ) 179 set @strTmp = @SqlSelect + ' * from ( ' + @SqlSelect + ' top ' + CAST ( @pageSize as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 180 + ' where ' + @ID + ' >(select max( ' + @ID + ' ) from( ' + @SqlSelect + ' top ' + CAST ( @pageSize * ( @page - 2 ) + @lastcount as Varchar ( 20 )) + ' ' + @ID + ' from ' + @tblName 181 + ' where (1=1) ' + @strCondition + ' order by ' + @fldSort + ' ' + @strSortType + ' ) AS TBMaxID) ' 182 + ' ' + @strCondition + ' order by ' + @fldSort + ' ' + @strSortType + ' ) AS TempTB ' + ' order by ' + @fldSort + ' ' + @strFSortType 183 else 184 set @strTmp = @SqlSelect + ' * from ( ' + @SqlSelect + ' top ' + CAST ( @pageSize as VARCHAR ( 4 )) + ' ' + @fldName + ' from ' + @tblName 185 + ' where ' + @ID + ' <(select min( ' + @ID + ' ) from( ' + @SqlSelect + ' top ' + CAST ( @pageSize * ( @page - 2 ) + @lastcount as Varchar ( 20 )) + ' ' + @ID + ' from ' + @tblName 186 + ' where (1=1) ' + @strCondition + ' order by ' + @fldSort + ' ' + @strSortType + ' ) AS TBMaxID) ' 187 + ' ' + @strCondition + ' order by ' + @fldSort + ' ' + @strSortType + ' ) AS TempTB ' + ' order by ' + @fldSort + ' ' + @strFSortType 188 end 189 end 190 191 -- ----返回查询结果----- 192 exec sp_executesql @strTmp 193 select datediff (ms, @timediff , getdate ()) as 耗时 194 -- print @strTmp 195 SET NOCOUNT OFF 196 GO 197
执行示例：exec proc_paged_2part_selectMax 'tb_testTable','ID,userName,userPWD,userEmail',10,100000,'ID',0,null,'ID',0
这种测试只在单机进行，并且没有在实际开发WEB项目中分页测试，测试项也比较单一，所以不够全面系统，但从其效率相比上，我们可以在数据库分页算法上进行有效的控制。
