Yehoshua Sagiv

The TSIMMIS Approach to Mediation: Data Models and Languages

TSIMMIS—The Stanford-IBM Manager of Multiple InformationSources—is a system for integrating information. It offers a datamodel and a common query language that are designed to support thecombining of information from many different sources. It also offerstools for generating automatically the components that are needed tobuild systems for integrating information. In this paper we shalldiscuss the principal architectural features and their rationale.

Answering queries using templates with binding patterns (extended abstract)

When integrating heterogeneous information resources, it is often the case that the source is rather limited in the kinds of queries it can answer. If a query is asked of the entire system, we have a new kind of optimization problem, in which we must try to express the given query in terms of the limited query templates that this source can answer. For the case of conjunctive queries, we show how to decide with a nondeterministic polynomial-time algorithm whether the given query can be answered. We then extend our results to allow arithmetic comparisons in the given query and in the templates.

Constraint checking with partial information

Constraints are a valuable tool for managing information across multiple databases, as well as for general purposes of assuring data integrity. However, efficient implementation of constraint checking is difficult. In this paper we explore techniques for assuring constraint satisfaction without performing a complete evaluation of the constraints. We consider methods that use only constraint definitions, methods that use constraints and updates, and methods that use constraints, updates, and “local” data.

Inference of inequality constraints in logic programs (extended abstracts)

Inequalities among sizes of arguments in logic programs are essential for showing termination and safety. “Termination” usually refers to termination of top-down evaluation (as in Prolog), and ‘(safet y“ refers to the property that bottom-up evaluation (using magic sets, for example) generates a finite answer. The approach toward logic programming enunciated by systems such as NAIL! [M*86, M*87] and LDL [NT88, Z88] puts the burden of showing termination and safety on the compiler. Therefore, there is a need to develop algorithms for these problems. Since safety and termination are undecidable for logic programs with function symbols [KiL88, Sh87], one can only hope for algorithms that check sufficient conditions that are strong enough to be practical.

Convergence of sideways query evaluation

Sets of Horn clauses with no function symbols and negation can be considered as a query language (sometimes called DATALOG) that generalizes relational algebra. For example, the following clauses define the transitive closure relation of a binary relation A (not definable in relational algebra): Rule 1: T(z, y) : -A(z, y). Rule 2: T(z, y) : -A(z, z),T(z, y). Here A is a database relation, (it does not appear on the left-hand side of any rule), whereas T is a non-database relation. A typical query would be T(5, CC)?, asking for all items reachable from item 5.

Optimizatioin of a Subclass of Conjunctive Queries

The optimization problem for a subclass of conjunctive queries which is formed by the union of the class of fan-out free queries and a subclass of typed fan-out queries is investigated. The typed fan-out queries in this class are obtained from simple tableaux by allowing at most one attribute to violate the simple-tableau property. The optimization problem for several restricted subsets of typed fan-out queries is already known to be NP-hard. It is shown that the queries under consideration possess several useful properties which are then used to obtain an O(n2) optimization algorithm based on the implication graph technique. The optimization of typed fan-out queries, obtained from simple tableaux by allowing at most two attributes to violate the simple tableau property, is shown to be NP-hard.

Optimization of a subclass of conjunctive queries

The optimization problem for a subclass of conjunctive queries which is formed by the union of the class of fan-out free queries and a subclass of typed fan-out queries is investigated. The typed fan-out queries in this class are obtained from simple tableaux by allowing atmost one attribute to violate the simple-tablau property. The optimization problem for several restricted subsets of typed fan-out queries is already known to be NP-hard. It is shown that the queries under consideration possess several useful properties which are then used to obtain an O(n2) optimization algorithm based on the implication graph technique. The optimization of typed fan-out queries, obtained from simple tableaux by allowing atmost two attributes to violate the simple tableau property, is shown to be NP-hard. The optimization of simple tableaux in the presence of functional dependencies is also investigated and is shown to be NP-hard.