Concept hierarchies for extensible databases

Abstract

This thesis presents a simple and efficient method of implementing a semantic type checking system for use with relational databases. Numeric data typically represent measures of a specific property or characteristic of a real world object. Computers manipulate only the numeric value. It is the responsibility of the user to ensure that the data are handled in a manner consistent with its meaning. If the semantics associated with the numbers are stored in a data dictionary, semantic consistency can be verified by the database system. This increases the integrity of data manipulation and helps ensure meaningful results. This thesis demonstrates a simple scheme of representing the property, or quantity, and unit of measure associated with numeric attributes. This information is then used to verify dimensional consistency of database queries and to automatically convert units across systems of measurement. Finally, a concept is defined for each relation in the database. These concepts can be used to build a concept hierarchy to help ensure queries are consistent with the semantics of the database design.