Debugging relational database queries with multiset constraints

Project facts

Project promoter
Complutense University of Madrid
Project Number:
ES07-0030
Target groups
Researchers or scientists
Status:
Completed
Initial project cost:
€5,500
Final project cost:
€5,247
From EEA Grants:
€ 5,247
The project is carried out in:
Spain

Description

Debugging is the process of finding errors in software systems and it is one of the most time consuming task in the software development process. This project aims at defining a new proposal for finding errors in queries over relational databases. The expected result is an integrated framework for detecting errors and debugging relational database systems. From the practical point of view, is expected to develop a working system which can be integrated with a Database System such as Oracle or MySQL. The ideas are expected to be published in relevant conferences and/or journals to share the obtained knowledge in order to produce future developments in this line. The donor partner, University of Oslo has a lot of experience in applying formal methods to real-life software problems, and in particular in the application of constraints, a key point of the project. Project promoter and donor partner will work together to implement a prototype, possibly in Java connecting to the Database System using JDBC, to check the assertions/trusted versions provided by the user with respect to the user queries.

Summary of project results

Debugging is the process of finding errors in software systems and it is one of the most time consuming task in the software development process. This project aimed at defining a new proposal for finding errors in queries over relational databases. Scientific activity and results performed during the stay of the Spanish researcher from Universidad Complutense de Madrid at Simula Research Laboratory in Norway are: 1) Design of a framework for debugging SQL series. Partners have developed a formal method for debugging large sets of related database queries in the relational database language SQL. The debugger proposes explanations to the user about erroneous information obtained in some query, which can be either “wrong information” that should not be part of the result or “missing information” that was expected but was not in the result. 2) Anonimity of medical databases. A joint work with the Cancer Registry Bureau at Oslo was performed. The purpose of this research was to increase the level of anonymity in medical databases with the aim to allow publishing data relevant for scientific research while avoiding the disclosure of private information.

Summary of bilateral results

The team at the University of Oslo / Simula Research Lab has a lot of experience in applying formal methods to real-life software problems, and in particular in the application of constraints, a key point of the project. Also, their experience in testing, a phase previous to debugging, contributes to define an integrated framework. Moreover, the Norwegian team has been working in projects dealing with Data Intensive Systems. The research from the University Complutense has experience in debugging databases using formal techniques. Therefore the collaboration has been mutually enriching and is leading to new interesting results. As a result of the stay, two scientific publications are in preparation: “Debugging SQL database queries by explanations”, and “Increasing anonymity in screening programs”. During the stay at Simula Research Laboratory, the Spanish researcher from Complutense University of Madrid established new collaborations with researchers at the host institutions other than the initial collaborator, as well as with the research team of the Norway Cancer Registry in Oslo, where among other activities a colloquium was held on September 2014. Partners forecast to continue the contact in order to complete the research initiated during the stay funded by NILS programme.