Optimizing Data Access

Abstract

An execution plan, as described in Chapter 6, is composed of several operations. The most commonly used operations are those that access, filter, and transform data. This chapter specifically deals with data access operations, or, in other words, how the database engine is able to access data. There are basically only two ways to locate data in a table. The first is to scan the whole table. The second is to do a lookup based on a redundant access structure (for example, an index) or on the structure of the table itself (for example, a hash cluster). In addition, in the case of partitioning, access might be restricted to a subset of partitions. This is no different from looking up specific information in this book. Either you read the whole book, you read a single chapter, or you use the index or table of contents to find out where the information you are looking is.