Corporate Information Factory

CHAPTER 9

The Exploration and Data Mining Data Warehouse Components

To understand the need for the exploration warehouse, consider the stance of the data warehouse administrator when faced with a large query—say a 72-hour query (as seen in Figure 9.1).

Figure 9.1  What happens when the 72-hour query attempts to execute against the data warehouse?

The normal data warehouse administrator takes a firm stand and does not allow the large query to execute against the data warehouse. The rationale the data warehouse administrator offers is plausible. If the data warehouse administrator allows a very large query to go into execution against the data warehouse, everyone will suffer from poor performance. The large query will drain huge amounts of resources, and the normal processing that all other users of the data warehouse have come to expect will not be possible. Therefore, the administrator tells the analyst she cannot use the data warehouse for very large queries.

Is this a reasonable stance? From the standpoint of the average user of the data warehouse, this is rational and reasonable. With that said, this stance is not reasonable from the standpoint of getting the full value out of the warehouse.

How the Explorer Handles Large Queries

There are legitimate users of the warehouse (typically called explorers) who submit large queries. Explorers are usually “out of the box” thinkers—they ...