Dissertation/Thesis Abstract

Satisfiability checking for quality assurance in relational data processing
by Heller, Kelly K., M.S., California State University, Long Beach, 2013, 234; 1524200
Abstract (Summary)

Developers using the SQL language lack relational SQL debugging tools. SQL also suffers from a lack of sufficient automated frameworks for unit testing and regression testing of queries. This thesis begins with an SQL bug taxonomy. The focus is on Core SQL as defined in the SQL:1999 standard. Features in Core SQL remain virtually unchanged through the latest standard, SQL:2011. Our bug taxonomy illustrates common coding defects related to NULL, LEFT JOIN, selectivity assumptions, and ordering assumptions. We highlight ways that silent failures can impact a databasedependent application. Our proposal for addressing SQL development errors is a verification tool (named POQ) that leverages annotated postconditions attached to each query statement. The formalism of domain relational calculus is the basis for axiomatizing queries. Subsequently, a theorem prover searches for counterexamples that demonstrate when the postcondition fails. Alternative projects using SMT solvers and static analysis methods are also discussed.

Indexing (document details)
Advisor: Gittleman, Arthur
Commitee: Ebert, Todd, Englert, Burkhard, Monge, Alvaro
School: California State University, Long Beach
Department: Computer Science
School Location: United States -- California
Source: MAI 52/03M(E), Masters Abstracts International
Subjects: Computer science
Keywords: Formal verification, Post conditions, Relational data, SQL bug taxonomy, SQL code defect, SQL language
Publication Number: 1524200
ISBN: 9781303521775
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy