Originally, this chapter was called "Reading Data Efficiently." Thinking about it for a while actually convinced me that "and Correctly" should be appended to the title. The reason is that this chapter is not just about speed; it is also about getting things right and ensuring that the result can be generated fast. What good is a fast result if it happens to be just plain wrong?
The following issues will be covered:
NULL handling
Fixing joins
Using cursors
Synchronized seq scans
Prepared queries
Indexes and foreign keys