添加聚集索引后高效分页查询的效率

本人使用的是sql 2012数据库

ALTER proc [dbo].[ex2](
    @count int --次数
)
as
begin
declare @i int
set @i = 0
while @i < @count
begin
SELECT * FROM product0   
    WHERE
    [bName] like  '%23%'
     Order by grade desc,createTime desc OFFSET 4 ROWS FETCH NEXT 3 ROWS ONLY ;
set @i = @i + 1
end
end

100次 用时82秒，通过为grade添加聚集索引，用时 84秒

GO
ALTER proc [dbo].[ex3](
    @count int --次数
)
as
begin
declare @i int
set @i = 0
while @i < @count
begin
SELECT * FROM product0 w1   
    WHERE id in   
        (  
            SELECT top 3 id FROM   
            (  
                SELECT top 7 ID, grade,createTime FROM product0 where [bName] like  '%23%' and
    [pName] like  '%%' and
    [find0] like  '%%' and
    [find1] like  '%%' ORDER BY grade desc
            ) w order by grade,createTime DESC
        ) order by grade desc,createTime DESC
set @i = @i + 1
end
end

100次 用时86秒，通过为grade添加聚集索引，用时 7秒

查询1000的后三条数据，效率比较

GO
ALTER proc [dbo].[ex1](
    @count int --次数
)
as
begin
declare @i int
set @i = 0
while @i < @count
begin
SELECT * FROM product0 w1, 
(
    SELECT TOP 3 id FROM 
    (
        SELECT TOP 1000 id, grade,createTime FROM product0
        where
            [bName] like  '%23%' and
            [pName] like  '%%' and
            [find0] like  '%%' and
            [find1] like  '%%' 
        ORDER BY grade DESC
    ) w ORDER BY grade ASC 
) w2 WHERE w1.id = w2.id ORDER BY w1.grade DESC, w1.createTime desc
set @i = @i + 1
end
end

100次，添加gradle聚集索引,用时9秒

GO
SET QUOTED_IDENTIFIER ON
GO
ALTER proc [dbo].[ex3](
    @count int --次数
)
as
begin
declare @i int
set @i = 0
while @i < @count
begin
SELECT * FROM product0 w1   
    WHERE id in   
        (  
            SELECT top 3 id FROM   
            (  
                SELECT top 1000 ID, grade,createTime FROM product0 where [bName] like  '%23%' and
    [pName] like  '%%' and
    [find0] like  '%%' and
    [find1] like  '%%' ORDER BY grade desc
            ) w order by grade
        ) order by grade desc,createTime DESC
set @i = @i + 1
end
end

100次，添加gradle聚集索引,用时9秒

GO
ALTER proc [dbo].[ex2](
    @count int --次数
)
as
begin
declare @i int
set @i = 0
while @i < @count
begin
SELECT * FROM  (
    SELECT TOP 3 * FROM (
        SELECT TOP 1000 * FROM product0
        where [bName] like  '%23%' and
    [pName] like  '%%' and
    [find0] like  '%%' and
    [find1] like  '%%'
         ORDER BY grade DESC 
         ) f 
        ORDER BY f.grade ASC
    ) s ORDER BY s.grade DESC ,s.createTime Desc
set @i = @i + 1
end
end

100次，添加gradle聚集索引,用时4秒

go
execute sp_helpindex @objname = 'product0'  --查看索引
go
 drop index product0.CLUSTER_id;  --删除索引
go
CREATE CLUSTERED INDEX CLUSTER_id ON product0(grade) --创建索引

本人通过添加和删除聚集索引发现：使用SQL server 2012提供的 OFFSET 4 ROWS FETCH NEXT 3 ROWS ONLY 的查询语句，在速率上没有明显的提示。