APPLYING SEMANTIC DATABASE PRINCIPLES IN A RELATIONAL ENVIRONMENT

摘要

SQL generally allows several solutions to one single problem. The reason for this is that SQL's language concepts lack orthogonality. Because more solutions are allowed, SQL users are uncertain whether the specific solution they chose is correct and performs well. This paper presents a method to tackle this practical problem by introducing the orthogonal semantic concepts from the Xplain data language in the development phase of relational databases. This conversion method results in unique, consistent, well-performing and portable SQL definitions and SQL queries. Over a period of several years, we tested the approach extensively and applied it in practice, for which we used databases and queries of different complexities. The method proved to be time saving in both design and implementation. This paper presents an overview of the principles used in the automated tool for converting safe Xplain into equally safe SQL.