Can i make this stored procedure more efficent?

I've got a sp i.e. below - which i'm looking to run anything from 0 to 1million times per hour, as such i'm wondering if theres any way i can make this more efficent?

E.g. Merge Call 2 and 3 maybe?
Create PROCEDURE [dbo].StoredProcedureA
 
	@CompanyId int,
	@UsersId int,
	@SentCount int out
 
AS
	 SET NOCOUNT ON
begin
 
------------------
-- CALL 1
------------------
Update Users 
set STATUSID=3 
WHERE UsersId = @UsersId
 
------------------
-- CALL 2
------------------
UPDATE [Company]
	SET 
		countA = countA+1
		,countB = countB+1
	WHERE CompanyId =@CompanyId
 
------------------
-- CALL 3
------------------
select top 1 @SentCount=countA 
from [Company] 
WHERE CompanyId =@CompanyId
 
end

Open in new window

paulCardiffAsked:
Who is Participating?
 
Aneesh RetnakaranDatabase AdministratorCommented:
Yup , you need a COMMIT TRANS statement at the end ...
In case of an error , the sp throws an error value of -201/-202/-203 ; and based on this return value you can set an appropriate error message

BEGIN TRAN

Update Users
set STATUSID=3
WHERE UsersId = @UsersId

IF @@ERROR<>0
BEGIN
      ROLLBACK TRAN
      SELECT -201 AS RetVal
      RETURN
END
------------------
-- CALL 2
------------------
UPDATE [Company]
        SET
                countA = countA+1
                ,countB = countB+1
        WHERE CompanyId =@CompanyId

IF @@ERROR<>0
BEGIN
      ROLLBACK TRAN
      SELECT -202 AS RetVal
      RETURN
END  
------------------
-- CALL 3
------------------
select top 1 @SentCount=countA
from [Company]
WHERE CompanyId =@CompanyId
IF @@ERROR<>0
BEGIN
      ROLLBACK TRAN
      SELECT -203 AS RetVal
      RETURN
END
 
CoMMIT TRAN
SELECT 1 AS RetVal --retval 1 means the operations were successfull

END
0
 
Aneesh RetnakaranDatabase AdministratorCommented:
i am not sure what exactly u mean by call 1 ,2 and 3
In case you need to perform all the 3 operations with each call, then you need to encapsulate this inside a transaction. Now, in case only one expression need to be executed with a call, then better u modify this as 3 separate sps.
also make sire that you have indexes on the columns which is there in the where clause
0
 
paulCardiffAuthor Commented:
Thanks for that
What would be the best way to put this in a single transaction ?
0
Free Tool: ZipGrep

ZipGrep is a utility that can list and search zip (.war, .ear, .jar, etc) archives for text patterns, without the need to extract the archive's contents.

One of a set of tools we're offering as a way to say thank you for being a part of the community.

 
Aneesh RetnakaranDatabase AdministratorCommented:

BEGIN TRAN 
 
Update Users 
set STATUSID=3 
WHERE UsersId = @UsersId
 
IF @@ERROR<>0 
BEGIN 
	ROLLBACK TRAN
	SELECT -201 AS RetVal
	RETURN
END 
------------------
-- CALL 2
------------------
UPDATE [Company]
        SET 
                countA = countA+1
                ,countB = countB+1
        WHERE CompanyId =@CompanyId
 
IF @@ERROR<>0 
BEGIN 
	ROLLBACK TRAN
	SELECT -202 AS RetVal
	RETURN
END  
------------------
-- CALL 3
------------------
select top 1 @SentCount=countA 
from [Company] 
WHERE CompanyId =@CompanyId
IF @@ERROR<>0 
BEGIN 
	ROLLBACK TRAN
	SELECT -203 AS RetVal
	RETURN
END 
 

Open in new window

0
 
paulCardiffAuthor Commented:
do i need to say comit at the end?
0
 
paulCardiffAuthor Commented:
sorry also what does " SELECT -202 AS RetVal" do?
0
 
PaultheBrokerCommented:
Check out this cool OUTPUT feature of MSSQL2005: it would prevent you having to rescan the whole of the company statement in CALL 3, so might be marginally more efficient if you are really worried about the processing time.....
http://msdn2.microsoft.com/en-us/library/ms177564.aspx

I think it would work something like this in your case
DECLARE @company table (countA int)
....(update statement here)
OUTPUT  inserted.countA INTO @company
SELECT @SentCount=FIRST(countA) FROM @company

The inserted table contains the value of whatever you've just inserted, so in this case it will be 'countA + 1'....
0
 
Ken SelviaRetiredCommented:
This is probably even faster;

Update Users
set STATUSID=3
WHERE UsersId = @UsersId
AND STATUSID <> 3   -- <-- Prevent updating if the value is already set

-- Get the value of @SentCount when we update the Company table
UPDATE [Company]
      SET
      @SentCount = countA = countA+1   -- <-- Here
            ,countB = countB+1
      WHERE CompanyId =@CompanyId
 

And TOP 1 really had no meaning here;

select top 1 @SentCount=countA
from [Company]
WHERE CompanyId =@CompanyId


I know what you are going to say;  "It means the FIRST matching row", but if there is more than one row, what is the FIRST without an ORDER BY?  I suspect there is only 1 matching row since CompanyID looks like the primary key of [Company].

Also for what it's worth in the UPDATE with OUTPUT, OUTPUT goes before the WHERE clause.  Since the accepted solution returns RetVal, he is now prepared to read the resultset of the proc.  In that case, eliminate the OUTPUT to a table variabale and just output CountA as the resultset (which is the default for OUTPUT anyway)  Still, the method I posted above is even better.

Experts - here is the challenge; How can he write both update statements in one batch to eliminate the need for BEGIN/END TRANSACTION?  I'm not sure it's possible.  Now that the SELECT from company is  elimianted, I'm not sure he needs the transaction anyway.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.