Using Delta Relations to Optimize Condition Evaluation in Active Databases

Abstract

We give a method for improving the effciency of condition evaluation during rule processing in active database systems. The method derives, from a rule condition, two new conditions that can be used in place of the original condition when a previous value (true or false) of the original condition is known. The derived conditions are generally more effcient to evaluate than the original condition because they are incremental--they replace references to entire database relations by references to delta relations, which typically are much smaller. Delta relations are accessible to rule conditions in almost all current active database systems, making this optimization broadly applicable. We describe an attribute grammar based approach that we have used to implement our condition rewriting technique.