Overview

This chapter examines the loading of data into an Oracle Database data warehouse. There are various ways in which the loading process can be made to perform better. This chapter will attempt to focus on the performance aspects of what is effectively a three-step process, and sometimes even a four-step process, including extraction, transportation, transformation, and loading processes. I like to add an extra definitional step to the loading process, called transportation. Transportation methods will also be discussed in this chapter, because some methods are better and faster than others and there are some very specific and highly efficient transportation methods specific to Oracle Database.

13.1 What Is Data Loading?

Data loading for a data warehouse is essentially the process by which data is loaded into that data warehouse. Unfortunately, loading data into a data warehouse is unlikely to ever be a simple process for an innumerable number of reasons; perhaps the most significant reason is that data originates from different sources and is thus likely to differ substantially. Therefore, the loading process can be split into a sequence of steps:

1. Extraction. This is the process of getting or pulling the data from a source. For example, queries against a legacy database can be used to generate flat text files.

2. Transportation. A step such as this is not always an explicit requirement, but often data must be transported from one machine to another, perhaps in a specific form such as a text file or Oracle transportable tablespace,...