TRANSLATING UPDATES OF RELATIONAL DATA BASE VIEWS

• Authors: S. S. Cosmadakis
• Venue: TRANSLATING UPDATES OF RELATIONAL DATA BASE VIEWS
• Year: 1983

We study the problem of translating updates of data base views. We disambiguate a view update by requiring that a specified view compliment (i.e. a second view which contains all the data base information omitted from the given view) remains constant during the translation. We study some of the computational problems related to the application of this general methodology in the context of relational data bases. We consider, for the most part, data bases consisting of a single relation, with functional dependencies as the only integrity constraints; we also restrict our attention to view defined by projections. We first give a characterization of complementary views (valid also in the presence of join dependencies), which leads to efficient algorithms for checking if two given views are complementary and for determining a non-redundant complement of a given view. We also show that the problem of finding a minimum complement of a given view is NP-complete. We then study in detail the problem of translating the insertion of a tuple into a view. We show how to do the translation in case the insertion is translatable, and we also develop a polynomial time algorithm for testing translatability; we also give two stronger, more efficient translatability tests. We show lower bounds for the complexity of the translatability problem, by proving that it becomes -hard if the view is given in an exponentially succinct way; an analogous result is shown for one of the stronger tests. We also examine the problem of determining a complement which renders a given insertion translatable; we find that it can be solved in time polynomial in the view, but becomes NP-hard if the view is given in an exponentially succinct way; again, analogous results are valid for the stronger tests. The above results are extended, in a straightforward way, to the cases of deletion and replacement of a tuple. Finally, we define and study a new kind of functional dependencies which is important in the context of complements, the explicit functional dependencies, (EFD''s), which intuitively state that some part of the data base information can be computed from the rest. We examine the interaction of EFD''s with functional dependencies and join dependencies, and we also extend our characterization of complementary views to allow for the presence of EFD''s.