Data Transformation Analysis Oracle Tips by Burleson Consulting

The Data Warehouse Development Life Cycle

DATA TRANSFORMATION ANALYSIS
Also recall that Guttbaum’s has a fact table in which each item appears as a separate row in the table. In this way, correlational analysis can be used to determine associations between grocery items. However, even though the database stores a row for each item in a transaction, there is no direct information about the different categories in which each item participates. A transaction may record the sale of two pounds of calf livers, but it does not record the fact that calf liver is a part of the beef category, that beef is a part of the meat category, and that meat is a part of the food category. Guttbaum’s classification hierarchy might look something like the hierarchy shown in Figure 3.18.

Figure 3.18 A classification hierarchy for Guttbaum’s Grocery.

In practice, just about every “object” in a data warehouse could become a part of a classification hierarchy, and because a data warehouse is subject-oriented, it is usually the subject that is classified. Another example would be a data warehouse for a bookstore where sales of books would need to be classified into fiction, non-fiction, and so on. The identification of these classification hierarchies is critical to warehouse analysis because a method must be designed to allow the subjects to be categorized for queries.

As noted earlier, the idea of classification hierarchies in a data warehouse analysis has parallels in the object-oriented world. In a C++ program, class hierarchies can be created to distinguish between different “types” of an object, where each type has its own data items and behaviors. The same principle applies to the data warehouse. The end user may want to query, How many paper products were sold last week? Unless our warehouse knows which items participate in the paper products categories, the warehouse will not be able to easily answer this query.
As we will discuss in the next chapter, classification hierarchies are added to basic entity/relation models so that the data warehouse analyst can describe each type of item. In Figure 3.19, you can see that a bill-of-material’s entity for an item has been extended to allow for different types of items. We’ll take a closer look at extending entity/relation models in the next chapter, when we begin to apply physical techniques to our logical systems analysis.

Figure 3.19 An extension of an entity/relation model.

---

This is an excerpt from "High Performance Data Warehousing".

If you like Oracle tuning, you may enjoy the book Oracle Tuning: The Definitive Reference , with over 900 pages of BC's favorite tuning tips & scripts.  You can buy it directly from the publisher and save 30%, and get instant access to the code depot of Oracle tuning scripts.