Custom paging query comparison Sql server 2000 , 2005 and 2012 (denali)

Custom paging is a very common requirement for any web application. It was always performance enhancer for our old applications. Linq and generic collections enabled us to process and query the result sets in memory. However, if the query can fetch only desired number of records with a better query performance, we need not perform any manipulation later on.

I am going to describe the custom paging query which had evolved with every Sql server upgrade. I will compare Sql Server 2000, 2005 and 2012 (denali) version to showcase the performance improvement. Sample DB isAdventureWorks.

Sql Server 2000: In 2000 we had to create a temp table before we skip certain number of records and fetch required number of records only. Temp table creation and inner join degrades the query performance here.


CREATE PROCEDURE GetPersonPaged_Sql2000
(
@startRowIndex int,
@pageSize int
)
AS
BEGIN
 SET NOCOUNT ON;

DECLARE @lastRowIndex int

SET @lastRowIndex = (@startRowIndex + @pageSize) - 1;

--Create a table variable
 DECLARE @TempItems TABLE
 (
 ID int IDENTITY,
 BusinessEntityID int,
 firstname varchar(200),
 MiddleName varchar(200)
 )
 -- Insert the rows from tblItems into the temp. table
 INSERT INTO @TempItems (BusinessEntityID,firstname,MiddleName)
 SELECT BusinessEntityID,firstname,MiddleName FROM Person.Person
 order by BusinessEntityID

-- Now, return the set of paged records
 SELECT p.* FROM @TempItems t
 INNER JOIN Person.Person p ON
 p.BusinessEntityID = t.BusinessEntityID
 WHERE ID BETWEEN @startRowIndex AND @lastRowIndex
 order by BusinessEntityID

END

Sql Server 2005: With Ranking functions introduced in 2005 upgrade it was easier to get a row number for each record (Please note that Sql Server never had any keyword RowNum like Oracle). ROW_NUMBER() is the function which helps sorting and numbering any result set. Please refer to http://msdn.microsoft.com/en-us/library/ms186734.aspx to know more about this function.
New implementation was even better using Common table expression CTE. Refer http://msdn.microsoft.com/en-us/library/ms190766.aspxto read more about CTE.

CREATE PROCEDURE GetPersonPaged_Sql2005
(
@startRowIndex int,
@pageSize int
)
AS
BEGIN
 SET NOCOUNT ON;
 DECLARE @lastRowIndex int

SET @lastRowIndex = (@startRowIndex + @pageSize) - 1;

;with CTE
 AS
 (
 select BusinessEntityID,firstname,MiddleName , ROW_NUMBER() OVER (order by BusinessEntityID ) as rownum
 from Person.Person
 )
 select * from CTE
 where rownum between @startRowIndex AND @lastRowIndex

END

It had improved the query performance by more than 80% than the Sql 2000 query. You can test the same by comparing execution plans.
Finally the winner is :

Sql Server 2012:It introduced new keywords like OFFSET and ROWS. It not only reduces the number of lines of code but improve the query performance by more than 50% when compared to Sql 2005 query using CTE.


CREATE PROCEDURE GetPersonPaged_Sql2012
(
@startRowIndex int,
@pageSize int
)
AS
BEGIN
 SET NOCOUNT ON;
 DECLARE @lastRowIndex int

SET @lastRowIndex = (@startRowIndex + @pageSize) - 1;

SELECT * FROM Person.Person
 ORDER BY BusinessEntityID
 OFFSET @startRowIndex ROWS
 FETCH NEXT @lastRowIndex ROWS ONLY;

END

Conclusion:    2012 query is the best of all three in terms of ease, lines of code and performance. If you are still using the old style query(even after upgrading to next version of sql server) for custom paging, this is high time you should switch to query with new syntax which is best suited for that particular version.

Note: I did not cover Sql server 2008 as it did not add any features which could add any value to custom paging querying.

I hope it is useful!

About these ads

Posted on May 3, 2012, in .NET, Sql Server and tagged , , , , , , , . Bookmark the permalink. Leave a comment.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 366 other followers

%d bloggers like this: