Online analytical processing (OLAP) is a widely used technology for facilitating decision support applications. In the paper, we consider partial aggregation queries, especially for partial top-k/bottom-k, which retrieve the top/bottom-k records among the specified cells of the given query. For the efficient processing of partial ranking queries, this paper proposes a set of algorithms using the RD-Tree, which is a data structure previously proposed for partial max/min queries. Through experiments with real data, we show the efficiency, robustness, and low storage overhead of the proposed method. (c) 2006 Published by Elsevier B.V.