Relational databases with set-valued attributes

Abstract

In this thesis, we investigate the use of sets in the relational model. In particular, we look at two separate appproaches. The first approach involves using collective sets as a collection of discrete data elements, while the second approach involves using disjunctive sets to represent incomplete information. We generalize these two approaches, proposing generalized sets, and consider a relational model which combines all three types of set — collective, disjunctive and generalized. Finally, we present a further generalization of the above approaches and introduce a relational model with restricted cardinality set. In this model, we explicitly store the minimum and maximum number of actual values in each set. This results in a model which is more simple, flexible, and enhances the richness of information that can be represented.