Abstract
Semantic Query Optimisation (SQO) is a considerably new approach to query optimisation, compared to the approaches used by the commercial databases. It takes the original query into its optimiser and analyses it by the use of automatically derived rules. From the answer and condition(s) of this query, the new rule(s) may be learned. The approach yields considerable time savings on query optimisation, especially when the query answer can be found from the rules. A main concern is that automatic rule derivation requires a long time in the SQO approach because of the database connection and its retrievals. To solve this problem, we apply mathematic logic to determine where a transition (from one rule to another) exists, and from here produce new rules. In this paper, two rule-mining modules useful in the SQO approach are Rule Transition and 'If and only If'. Computational results of the modules are very promising for learning new rules, and these rules improve the query executions.