Question

I have a dynamic query that runs indentifying CDs that members have not rented yet. I am using the NOT IN subquery but when I have large member table it makes them really slow. Any suggestions how to optimize the query

SELECT DVDTitle AS "DVD Title" 
FROM DVD 
WHERE DVDId NOT IN 
    (SELECT DISTINCT DVDId FROM Rental WHERE MemberId = AL240);

thanks

Answer 1

Using NOT EXISTS will have slightly better performance because it can “short circuit” rather than evaluating the entire set for each match. At the very least, it will be “no worse” than NOT IN or an OUTER JOIN, though there are exceptions to every rule. Here is how I would write this query:

SELECT DVDTitle AS [DVD Title] 
  FROM dbo.DVD AS d
  WHERE NOT EXISTS 
  (
    SELECT 1 FROM dbo.Rental 
     WHERE MemberId = 'AL240'
     AND DVDId = d.DVDId
  );

I would guess you will optimize performance better by investigating the execution plan and ensuring that your indexes are best suited for this query (without causing negative impact to other parts of your workload).

Also see Should I use NOT IN, OUTER APPLY, LEFT OUTER JOIN, EXCEPT, or NOT EXISTS?