Title: Well-Founded Views in Constraint Databases: Incremental Materilization and Maintenance
Authors: James J. Lu, Bertram Ludascher, Joachim Schue, V.S. Subrahmanian
Abstract
In this paper, we study the problem of incremental maintenance of materialized views when the view definitions are allowed to contain: (a) constraints, and (b) well-founded semantics of negation. A materialized view in our framework consists of a set of constrained non-ground tuples. We develop incremental algorithms to compute such non-ground materialized well-founded views, extending previous results of Gottlob et. al. and Either et. al. With respect to (b), previous attempts at view maintenance have been limited to the case when stratified negation is used. In contrast, in this paper, we allow arbitrary (stratified and/or unstratified) usage of negation. Recursive views are also allowed. We develop a view maintenance algorithm called WF-Red for handling updates in such cases. WF-Red takes to parameters as input - a list of insertions and a list of deletions. to handle insertions, the list of deletions is kept empty and vice-versa. The algorithm correctly handles both insertions and deletions to the materialized view.