Database Performance Tuning and Optimization

Rating:
Summary: Disappointing and out of date
Review: I didn't really order this book expecting a practitioner's guide to Oracle. It's written by a respected and widely published academician and I was looking forward to a melding of theory and practise. All I found was a book full of filler (lots of queries and printouts when summaries would have done) that is way out of date. For a performance book published in 2003 to not even mention Statspack (while devoting a chapter to bstat/estat) tells me that the author stopped working on Oracle at 8.0 and has not caught up. I have to admit that I cannot write a thorough, page by page review of this book because it is packed up for return to Amazon (a first for me). But to give you a couple of examples of the shallow treatment of the subject in this book: the index does not list the term histogram (poor editing) and the section devoted to it does not even mention that there are problems getting the optimizer to make full use of histograms when bind variables are used (though researchers have shown that histograms are not fully ignored when using binds as has been published many times). The section on partitioning doesn't go into the advantages and drawbacks of global vs. local indexes on partitions and doesn't in fact go into any substantive treatment of partitioning options at all. This book is over-priced and under-powered. I gave it two stars instead of one simply because someone who is a student trying to transition from pure theory into practise may find the reasurring presences of familiar terminology and formatting helpful. There is some material on the mathematical foundation of the relational model that could also be helpful to the student and practitioner in the field. But if you are looking for real performance books you would be far better off looking at those from authors of the Oak Table (search for Oak Table on Google to find this), such as Tom Kyte and Jonathan Lewis.