Category: Query Tuning

Erik Darling hits one of my favorite topics:

The first rule of rewrites is that they have to produce the same results, of course. Logical equivalency is tough.

In today and tomorrow’s posts I’m going to compare a couple different scenarios to get the top value.

There are additional ways to rewrite queries like this, of course, but I’m going to show you the most common anti-pattern I see, and the most common solution that tends to work better.

Click through to see when each works better.

Itzik Ben-Gan continues reviewing solutions to a tricky problem:

Last month I covered a solution based on interval intersections, using a classic predicate-based interval intersection test. I’ll refer to that solution as classic intersections. The classic interval intersections approach results in a plan with quadratic scaling (N^2). I demonstrated its poor performance against sample inputs ranging from 100K to 400K rows. It took the solution 931 seconds to complete against the 400K-row input! This month I’ll start by briefly reminding you of last month’s solution and why it scales and performs so badly. I’ll then introduce an approach based on a revision to the interval intersection test. This approach was used by Luca, Kamil, and possibly also Daniel, and it enables a solution with much better performance and scaling. I’ll refer to that solution as revised intersections.

Read on for one class of solution which performed quite well.

Itzik Ben-Gan has some solutions to show:

This month, I’m going to start exploring the submitted solutions, roughly, going from the worse performing to the best performing ones. Why even bother with the bad performing ones? Because you can still learn a lot from them; for example, by identifying anti-patterns. Indeed, the first attempt at solving this challenge for many people, including myself and Peter, is based on an interval intersection concept. It so happens that the classic predicate-based technique for identifying interval intersection has poor performance since there’s no good indexing scheme to support it. This article is dedicated to this poor performing approach. Despite the poor performance, working on the solution is an interesting exercise. It requires practicing the skill of modeling the problem in a way that lends itself to set-based treatment. It is also interesting to identify the reason for the bad performance, making it easier to avoid the anti-pattern in the future. Keep in mind, this solution is just the starting point.

Click through for a solution which is straightforward but slow.

Joe Obbish has a solution:

Itzik Ben-Gan posted an interesting T-SQL challenge on SQL performance dot com. I’m writing up my solution in my own blog post because I have a lot to say and getting code formatting right can be tricky in blog post comments. For reference, my test machine is using SQL Server 2019 CU14 with an i7-9700K CPU @ 3.60 GHz processor. The baseline cursor solution completes in 8465 ms on my machine.

Click through for a bit of formal logic and a lot of tuning.