Objectives:
This presentation is based on real life query examples. It will take the audience on a trip through different ideas and tiny tricks that we have used in our shop to deal with special situations, hard-to-explain optimizer behaviour and “unsolvable” performance issues. It will go through examples of performance issues and ways we solved them – covering issues caused by unexpected data type conversions or lack of them, issues caused by padded or not padded indexes (why don’t they work?!) and issues caused by joining multiple tables on range predicates (and MQTs that saved us).
Outline:
Example
Example
Example
Example
Example
Full Abstract:
Query optimization today is hardly what it used to be. Queries are no longer comprised of two or three equal predicates. With today’s query generators and new (outsourced) applications comming in, query optimization is becoming highly valued skill. While trying to find their way through the mess of joins, subqueries and all the new functionality that comes in with every new DB2 version, DBAs could often use some tips and tricks to make their lives easier. This presentation offers not really ready to use cookbook, but rather gives ideas that could help DBAs to solve their performance issues.