ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES".
No. 8. Vol. 26. 2020

DOI: 10.17587/it.26.460-471

A. N. Rodionov, Dr. of Tech. Sc., Computer Centre of Far-Eastern Branch of RAS, e-mail: ran@newmail.ru

Conceptual and Logical Data Modeling: Detection and Configuration of R-types and n-ary Relationships in Domains

The paper addresses the problems of detecting and configuring complexes of high-level database relationships (relationships with arity 3 and above) and representing part of the latter by means of P-types. Ð-type is one of the categories of synthetic object types whose instances are associated with linlcs that occur between objects. Such types can come with their own properties and can interact with other objects. It is pointed out that multi-valued functional dependencies, which can be present among the foreign key attributes of header document types, uniquely identify both the n-ary relations themselves and the corresponding P-types. P-types that replace n-ary relations eliminates the occurrence of tuple update anomalies in the header documentary types. We study the influence of different relationship cardinality constraints that involve multiple entity types on the attribute composition of the corresponding P-types. It is introduced the concept of a binary relations hipscomplex that subsume initial interacting types. On the basis of the proposed p/a criteria for binary relations, which allows to distinguish potential and actual relations of a complex, it is developed a method for ascertaining primary relations that are subject to storage, and secondary relations that derive from primary ones. It is concluded that all high-level p-relations can be deduced (by means of natural join operation) from potential binary relations, and all binary a-relations can be obtained by applying projection operations to some source n-ary a-relation, if one exists in the domain. In conclusion, an example of linked P-types that may appear in entity type clusters is given, and current modeling problems are listed.
Keywords: high-level relationships, P-types, p/a-binary relations, multi-valued dependences, ÐSI entity cluster

To the contents