Look ere ye leap.
To convert the art of SQL tuning to a science requires a common language, a common paradigm for describing and solving SQL tuning problems. This book teaches, for the first time in print in any detail, a method that has served me well and served others I have taught over many years. I call this method the query diagramming method.
Like any new tool, the query diagramming method requires some up-front investment from the would-be tool user. However, mastery of this tool offers tremendous rewards, so I urge you to be patient; the method seems hard only for a while. Soon, it will lead you to answers you would never have found without the tool, with moderate effort, and in the end it can become so second-nature that (like the best tools) you forget you are using it.
Since I am asking for your patience, I begin with a discussion of why this tool is needed. Why not use a tool you already know, like SQL, for solving performance problems? The biggest problem with using SQL for tuning is that it presents both too much and not enough information to solve the tuning problem. SQL exists to describe, functionally, which columns and rows an application needs from which tables, matched on which join conditions, returned in which order. However, most of this information is wholly irrelevant to tuning a query. On the other hand, information that is relevant, essential even, to tuning ...