Managing Reference Data in Enterprise Databases (The Morgan Kaufmann Series in Data Management Systems)

Managing Reference Data in Enterprise Databases explores technologies and techniques for managing data that's used to categorize or sort data, with emphasis on keeping it clean, consistent, and current. Reference data, in the lingo of author Malcolm Chisholm, is primarily any data that is used to categorize other data. The U.S. Postal Service abbreviations "PA" for Pennsylvania and "HI" for Hawaii are examples of reference data. Under Chisholm's definition, reference data is generally not data that exists on a remote system, or that another organization maintains. This book has an academic tone and steers clear of implementation issues that exist in particular database management systems (DBMS); it's a platform-independent engineer's handbook, full of structural diagrams, standard SQL queries, and pseudocode.

Using the clear and widely understood IDE1FIX means of graphically representing table relationships, the author shows how reference tables exist alongside other database contents. He presents some of the reference data problems that pop up in real life (in one succinct example, he writes about getting bonds rated "AAA" to come ahead of bonds rated "A," even though "A" comes first in ASCII sorts). His content will make you think about how to redesign your databases to use reference data more reliably, and offer you enough advice to enable you to realize your designs. --David Wall

Topics covered: Efficient database design as it relates to reference data, or data that's used to categorize other pieces of data. Sections deal with establishing table relationships, deciding on special values (such as those for "none" versus "not applicable"), and dealing with multiple languages. There's also information on referring to remotely maintained reference data, such as abbreviations lists maintained by the International Standards Organization (ISO).