The importance of a good browser
Ralph Kimball, et al.
Creating a data warehouse requires that you bring many components together, like a database, a query tool and a reporting device. This tip from Ralph Kimball, Laura Reeves, Margy Ross, and Warren Thornthwaite's book,
The data warehouse marketplace has been using the word browsing for at least fifteen years. In this context, it means interactively examining the relationships among attributes in a dimension table. Data warehouse browsing has nothing to do with browsing the Internet. We will use the term browsing in the older data warehouse sense, although we realize our claim to this word is a lost cause.
A good browser should allow the user to explore an unfamiliar dimension table and reveal the relationships in it. The user should be able to enumerate all the distinct values of an attribute as a user interface action that takes place within one or two seconds. There should be a threshold number of distinct values above which the browser simply returns the COUNT DISTINCT rather than each distinct value. The results of a browse request must be affected by other constraints that may have been set in the dimension table during the browse session. Finally, a set of constraints that have been carefully specified during a browse session must be able to be saved with a user defined name so that this set of constraints can be used at various times in the future in queries and reports. Please ask your favorite query tool vendor for this set of features. If you get a blank stare, ask why the vendor's developers don't spend time building queries against real dimension tables with their customers.
Click on the title to learn more about The Data Warehouse Lifecycle Toolkit.
Are there data warehousing issues you want to read about? Email and let us know what is on your mind.
This was first published in August 2001