Ph.D. Thesis

Vladimir Toncar. Datalog Extensions and the Use of Datalog for Large Data Analysis. Ph.D. Thesis, University of West Bohemia, 2000.


This thesis is centered around Experimental Deductive Database (EDD), a deductive system that is being developed at the University of West Bohemia. The EDD system is built on top of an SQL-based relational database. It uses Datalog as its programming language and is based on the translation of Datalog rules to SQL.

The aim of the thesis is to propose useful extensions of the Datalog language and to study the possibilities of using the EDD system for the analysis of large data.

We study three important Datalog extensions. First of all, we deal with the use of imperative programming in Datalog programs. We study the influence of imperative instructions (e.g. assignment) on logic rules, classify the rules according to the influence and propose a translation algorithm for each class of rules. The second extension that is studied in this thesis is negation. The stratified negation semantics is used and a translation algorithm for stratified programs is proposed. Our last studied extension is aggregation. We present a translation algorithm and semantics for predicates with aggregates.

In our research of the use of Datalog in data analysis, we focus on OLAP and Data Mining. We propose an approach to implementing OLAP operations in Datalog and after that we present a Datalog application for mining association rules.