Martin-Luther-University Halle-Wittenberg
Computer Science Institute
Databases Group

SQLLint - Detecting Semantic Errors in SQL Queries

Stefan, Brass, Christian Goldberg


Project Goal

We investigate classes of SQL queries which are syntactically correct, but certainly not intended, no matter for which task the query was written. For instance, queries that are contradictory, i.e. always return the empty set, are quite often written in exams of database courses. Current database management systems, e.g. Oracle, execute such queries without any warning. In this project, we explain serveral classes of such errors, and develop algorithms for detecting them. Of course, questions like the satisfiability are in general undecidable, but our algorithm can treat a significant subset of SQL queries. We believe that future database management systems will perform such checks and that the generated warnings will help to develop code with fewer bugs in less time.


Project Participants


Papers


Talks


Exam Evaluations


Software


CG, Halle a.d. Saale, 29.08.2009; 15:53:09

Valid HTML 4.01! Valid CSS!