Most queries are just this easy to tune, once you master the method
this book describes. Usually, a missing index or some trivial problem
in the SQL is the only thing obstructing the optimizer from
delivering the optimum execution plan you choose, or a plan so close
to optimum as not to matter. You rarely need the elaborate special
techniques covered at the end of Chapter 6 and
throughout Chapter 7.
However, when you do need the special-case techniques, you
really need them!
The primary value of the method is that it leads you quickly to a
single answer you can be completely confident in, without any nagging
worries that long trial and error might just lead you to something
better. When the method leads you to a super-fast query, you find
little argument. When the method leads to a slower result than
you'd like (usually for a query that returns
thousands of rows), you need to know that the slower result really is
the best you can do without stepping outside the SQL-tuning box. The
outside-the-box solutions for these slower queries tend to be
inconvenient. It's invaluable to know with complete
confidence when these inconvenient solutions are truly necessary. You
need to justify this confidence without endless, futile attempts to
tune the original SQL by trial and error, and with solid arguments to
make your case for more difficult solutions when needed.