Resolving the query inference problem by optimizing query formulation cost

The query inference problem in relational databases consists in translating a sentence of a query language into an unambiguous representation of the query. We describe a new approach for solving the query inference problem based on query formulation cost optimization. To this end the logical schema of a database is modelled as a graph, where each node is a relation, and each arc is an inter-relation link. A relation may be chosen so as to define a user viewpoint, i.Êe. a perspective for accessing data. From a given viewpoint, a virtual derived relation is built, on the basis of a heuristic method which minimizes the cost for query formulation by considering semantic aspects of the inter-relation links. The derived relation is determined by computing the spanning arborescence with maximum semantic content on the database graph. The user viewpoint technique is presented with reference to the SHADOW RDBMS and its advantages are pointed out in terms of facilities in the application software design and development.