Category: Query Tuning

Paul White explains spill levels:

At this point, you might be wondering what combination of tiny memory grant and enormous data size could possibly result in a level 15,000 sort spill. Trying to sort the entire Internet in 1MB of memory? Possibly, but that is way too hard to demo. To be honest, I have no idea if such a genuinely high spill level is even possible in SQL Server. The goal here (a cheat, for sure) is to get SQL Server to report a level 15,000 spill.

The key ingredient is partitioning. Since SQL Server 2012, we have been allowed a (convenient) maximum of 15,000 partitions per object (support for 15,000 partitions is also available on 2008 SP2 and 2008 R2 SP1, but you have to enable it manually per database, and be aware of all the caveats).

Read the whole thing.

Richie Rump shows how to limit the number of columns returned in an Entity Framework query:

This one’s a bit more tricky but let’s walk through it. We’re getting data from the Posts table where the Tags column equals “<sql-server>” and selecting every column from both the Posts and PostTags tables. We can tell because there are no specified properties in the Select. Even though this statement looks more complex it’s only three lines and looks somewhat like a SQL statement. But it’s really a LINQ (Language Integrated Query) statement, specifically a LINQ to Entities statement. This LINQ statement will be translated into this SQL statement:

Read the whole thing.

Brent Ozar introduces a new website:

It’s a free community service. We hope you love it as much as we do, especially all the cool new execution plan icons by our illustrator, Eric Larsen.

Down the road, we’re thinking about adding logins (so you can see past plans you’ve submitted), execution plan advice, image and HTML downloads (so you can embed the plan in your own blog or report), zooming, and more.

Check out the Paste The Plan website.

Greg Gonzalez with the announcement of the week:

Which brings me to today, one of the most exciting days in the history of SQL Sentry. I would like to introduce to you the new unified Plan Explorer v3! This edition includes all features of Plan Explorer PRO, plus two new and incredibly powerful features: Index Analysis and Performance Profiling.

These new features were originally going to be part of yet a 3rd edition called Plan Explorer ULTIMATE, and at an even higher price point. However, we ultimately came to our senses and realized that this would be counter to the mission, and that the only way ahead was to make all features of Plan Explorer freely available to everyone.

Andy Mallon goes in-depth with two of the new features:

If you’ve used Live Query Profile in Plan Explorer (or Live Query Statistics in SSMS), then you’re familiar with seeing your query plan play live as it executes. Performance Profiling adds some special sauce to that feature. Its like a DVR for the live query profile, and also displays the performance metrics as a graph below the plan diagram. The graph gives performance metrics by time, rather than by operator. In complex plans, you might have many operators doing work at the same time, and you’ll be able to see that better by looking at the graphs. Adding a time axis to the plan is a pretty ingenious way to give a new way to look at performance.

You can then replay the “DVRed” query, and re-watch the profile without re-running the query. If you save the plan as a .pesession file, the DVR’ed query with performance profile is included in that file. You can even change the speed to watch in slow-mo or fast-forward.

If you don’t already use Plan Explorer, you’re missing out on the single best query tuning add-in.

Kendra Little describes the nested loop operator in execution plans:

One of the first things we often learn when we’re looking at plans is that SQL Server uses estimates. And sometimes, those estimates are wrong. At first glance, this looks really wrong– it estimated 11.5 rows, and actually got 20,825 rows!

Read the whole thing to see how nested loop estimates which seem way off might not actually be.

Brent Ozar has a guide on troubleshooting parameter sniffing:

Parameter sniffing fixes are based on your career progression with databases, and they go like this:

1. Reboot the server! – Junior folks panic and freak out, and just restart the server. Sure enough, that erases all cached execution plans. As soon as the box comes back up, they run rpt_Sales for China because that’s the one that was having problems. Because it’s called first, it gets a great plan for big data – and the junior admin believes they’ve fixed the problem.

2. Restart the SQL Server instance – Eventually, as these folks’ careers progress, they realize they can’t go rebooting Windows all the time, so they try this instead. It has the same effect.

If a reboot can’t fix the problem, I’m out of ideas…

By the way, I second Brent’s recommendation of Erland’s query plan article.  Erland doesn’t publish frequently, but when he does it’s worth the wait.