Efficient rank join with aggregation constraints

Abstract

We show aggregation constraints that naturally arise in several applications can enrich the semantics of rank join queries, by allowing users to impose their application-specific preferences in a declarative way. By analyzing the properties of aggregation constraints, we develop efficient deterministic and probabilistic algorithms which can push the aggregation constraints inside the rank join framework. Through extensive experiments on various datasets, we show that in many cases our proposed algorithms can significantly outperform the naive approach of applying the state-of-the-art rank join algorithm followed by post-filtering to discard results violating the constraints.