Columnists Commentary

Within a database, a table’s meaning often becomes ambiguous.  A table design may start out clear but can then get cluttered with columns that may or may not be needed.  Designers should think about tables not as an arbitrary collection of data points but as the set of known truths about an entity.  A table containing customer data contains the known truth about the customers.  By understanding the validity of table structures as the known truth about an entity, rather than a generalized dumping ground for data somehow related to a generalized subject area, database designers can better determine if changes or additions will stretch the table beyond its intended scope.