PASS Summit 2014 is here!! Let the whole week of learning and networking begin! Kicking off my week with an awesome event by Red Gate software called “SQL in The City”. There were plenty of great sessions on Continuous Integration, Source Control, and other subjects. One of the session I attended was “How to Read an Execution Plan” by Grant Fritchey (@GFritchey). There were several things I learned in that session:
- Always look at the first operator (SELECT, INSERT, UPDATE, DELETE)
- Go to properties of the first operator and check for “Reason for Early Termination” (Good Enough Plan, or timeout, etc)
- You could also get the Query hash and the Query Plan Hash of the statement
- Check for Warnings
- Check for most costly operation
- Caveat: Just because it’s the most costly operation does not mean it is the issue
- Understand what the operator are doing. Hint: put your mouse over the operator and read the tool tip. It provide a description of what the operator is trying to do
- Check for fat pipes
- Pipes representing the amount of rows the operator is passing to the next operator
- According to Grant, if you see an unusual flow of pipes such as a fat pipe, follow by a small pipe, follow by a fat pipe, please call him. Something is terribly wrong 🙂
- Check the extra operations
- Example: Sort operation, Compute Scalar operation.
- Check for Scans.
Like anything in SQL Server, the mantra “It Depends!” still is applied here. These are generalize guidelines on identifying tuning opportunity when looking at execution plans. Caveat: These are just several things that I remembered from the session. I would definitely recommend everyone to attend this event in the future. To find out more about this event, check out SQL in The City website at http://sqlinthecity.red-gate.com.
I ended the day at a networking event called “BI over Beers” at the Tap House. These are what this week is all about, improving your technical skills and growing your network as a professional. Who knows, you could ended with few good friends. If you’re like me, you’ll make lots of ’em.