Interval hierarchies and their application to predicate files

Abstract

Predicates are used extensively in modern database systems for purposes ranging from user specification of associative accesses to data, to user-invisible system control functions such as concurrency control and data distribution. Collections of predicates, or predicate files, must be maintained and accessed efficiently. This paper describes a dynamic index, called an interval hierarchy, which supports several important retrieval operations on files of simple conjunctive predicates. Search and maintenance algorithms for interval hierarchies are given. For a file of n predicates, typical of the kind expected in practice, these algorithms require time equal to Ο (log n ).