And now for something not so different with week 6 of the Expert Performance Indexing for SQL Server book giveaway. So far, I’ve sent out five copies of the book based on comments from these (1, 2, 3, 4, and 5) posts.
Expert Performance Indexing for SQL Server is a book I co-wrote with Grant Fritchey (Blog | @gfritchey) last summer, which can be a valuable resource to anyone building indexes on SQL Server. As part of finishing the book, Apress sent me a number of copies of the book to share with people. I figured the best way to share these out is to give them away to people that comment on this blog.
So here’s a topic for leaving a comment this week…
The Question
For week six, the question will focus on chapters eight, which covers index myths and best practices. The topic for discussion in the comments this time is…
What indexing myths have you had to deal with? Have you had trouble with those who think databases don’t need indexes? Have you dealt with those that believe key column order doesn’t matter? Is there one you’ve dealt with that isn’t often talked about? Or how about an indexing best practice that you use?
If you have a story to share in this area, leave it in a comment below. After a week, I’ll select one of the comments and send that author a copy of the book.
August 1 Update
A bit late, again, on selecting the winner for this post, but after looking over the comments, the winner for the book this week is Peter Schott. Peter worked with the “order doesn’t matter” myth. This is probably one of my favorite myths, mainly because it logically doesn’t make sense that order would matter. Order does indeed matter for key columns and the selectivity or frequency of filtering on those columns makes a huge difference. This myth is dealt with directly in chapter 8. Learn more about indexing myths and best practices in Expert Performance Indexing for SQL Server.
The other comments this week were from:
- Aaron dealt with a problem of heaps and compression and how their maintenace appeared to interrupt the desired state
- Erik dealt with primary key columns needing to be added added as key columns to non-clustered indexes
- John had a client that had a ‘no downtime’ requirement and bad practices around index and statistic maintenance.
