NYOUG 2017 Fall General Meeting: Guesswork and the Optimizer
Guesswork and the Optimizer
The presentation will examine assumptions and guesses related to computing selectivity that the Oracle optimizer has to make in order to be able to optimize SQL statements. The notion of “confidence” of cardinality estimates, a measure of the level of guesswork used by the optimizer, will be introduced. The metric will be used to explain some of the challenges the Oracle optimizer faces.Finally, the presentation will focus on the types of SQL predicates that increase the likelihood of suboptimal
Finally, the presentation will focus on the types of SQL predicates that increase the likelihood of suboptimal execution plans. Multiple ways to prevent or mitigate those kinds of problems will be discussed in detail.
Iordan Iotzov is the Director of Data&Systems Administration at News America Marketing (NewsCorp) with more than fifteen years of administration and development experience. He has a MS degree in Computer Science, holds numerous certifications, and has worked on complex problems across various industries, providing original and effective solutions. Iordan has presented at leading database conferences and actively blogs at http://iiotzov.wordpress.com/.