On the foundations of the universal relation model

From MaRDI portal





The universal relation model aims at achieving greater access path independence in databases than the pure relational model yields. Assumptions underlying the universal relation model are given, mainly the uniqueness of roles of attribute names, and of their basic relationships. Two different approaches to universal relations are given: 1. A universal relation is defined by the user a priori and serves as the base for his queries. Either there exists only a unique universal relation and the database can be wiewed as its physical representation, or weak universal relations (weak instances) are used. The concept of representative instance as intersection of all weak instances is given. 2. The user queries about any set of attributes without noticing any relationship, and a universal relation based on user queries and data dependencies is constructed algorithmically a posteriori. As basic algorithm a union of projections of lossless joins is used and considerably generalised in the paper. The relationship between these two approaches is investigated.




Cited in
(45)








This page was built for publication: On the foundations of the universal relation model

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3677202)