asked on

Best solution to SQL insert-retrieve query function. (Tacobell777 - but others too!)

Been looking forever for a better solution to this than wrapping two queries into a transaction in order to retrieve the new record ID of a row that was just inserted by the query. Don't want to have to write stored procs for every query I use and hate having to hit SQL twice to just get my new records auto-increment ID value. This is primarily for TacoBell777, but if anyone else has any really nifty <CFQUERY> tricks or tips here I'll possibly split some points their way too!

BTW - for easy searching by others and for my own referrence, TacoBell777 if you could put a complete example of a short insert-retrieve function here in addition to your other example I'd love it! Lets say that table TBLMYDATA has the following columns:
RowID (INT, AUTO, PK)
TextCol (NVARCHAR)
NumCol (INT)
BitCol (BIT)

Thanks again for the cool tip and for understanding SQL much better than I do. :)

SOLUTION

reggi635

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

digicidal

ASKER

Hmmm... I was under the (apparently mistaken) belief that you could only guarantee that @@identity would return the correct ID when run in a storedproc - since @@identity is a global and since SQL is a multithreaded app as is CF that you might possibly have someone do an insert on the server in the brief space between the INSERT and the SELECT statements. In which case the @@Identity would actually return the recordID of the second users INSERT and not the first.

Is this incorrect? Tacobells solution is the following:

<cfquery ...>
DECLARE @pk INT

INSERT statement here

SET @pk = SCOPE_IDENTITY()

SELECT @pk AS primaryKey
</cfquery>

You actually referrenced that is the last part of your comment... can you (or someone else for that matter) provide evidence or links to conclusively support one method or the other in most cases (I realize that other steps may need to be taken in the case of really complex transactions or triggers on the server side). You've earned some points already reggi635 - and if Tacobell doesn't choose to comment here I'll award you instead with the whole thing. I'm really just looking for a discussion on which method is truly better and why. Particularly in a MS SQL CF MX situation. I'm just convinced that my old method (transactions with a select MAX(ID) is definitely NOT the best, but I would love to hear some evidence as to why a SCOPE_IDENTITY() method is preferrable or not to @@Identity.

Tacobell777

yes you are better of using SCOPE_IDENTITY()

you guys are robbing my points!!!! ;-))

ASKER CERTIFIED SOLUTION

Tacobell777

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

SOLUTION

Tacobell777

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial