Why Brent Ozar Wants You to Think Like a SQL Server Engine

Hey, DBAs – are your SQL Server Execution Plans sucking? 

 

If you answered affirmatively with a deep sigh, then you missed the trick question. Yes, of course they are! They’re “sucking” data from various tables or indexes to complete your query. Since we know how much time you likely spend a day on developing queries, we want to give you the resources to help your Execution Plans work faster and more efficiently ― so you can do the same.  

 

We asked everyone’s favorite Microsoft Certified Master, MVP and SQL Server whisperer, Brent Ozar, to show how SQL Server engines read and follow Execution Plans. In a webcast for DBAs who know how to write queries but could use some help speeding them up and reducing costs, Ozar walks through SQL Server best practices.

 

Want a preview? Here are a few quick tips:

 

  • Say no to sorting – Sorting in the database adds time and cost to a query by sorting all data first, not just the data you want to pull. If sorting is unavoidable, at least minimize the fields involved.
  • Indexes are your friends – An index copies the table you want to query, stored in the order you want with the fields you need. In one example from the webcast, Ozar reveals how he executed a query 4,000 times faster after fixing the index.
  • Tune the indexes – To fine-tune SQL Server performance, simplify the index. The best practice is to start with five or fewer indexes with five or fewer tables each.
  • Evaluate your non-clustered indexes – Remember, narrow indexes take less space, but wider indexes cover more queries. What’s most important to your end user?
  • Consider statistics – To build better plans, analyze how plans perform. Use statistics to calculate the memory required for a query and adjust accordingly.

 

In the webcast, Ozar plays the role of the end user. You’ll learn to think like a SQL Server engine by manually running SQL Server Execution Plans using Ozar’s data sets.

So if you’re ready to take your queries to the next level, download the associated PDF of data sets and check out the on-demand webcast today. 

 

Anonymous