Abstract: ROLAP is a kind of OLAP（Online Analytical Processing） in which it is widely used. Its major function is to manage aggregate data which is needed in the decision-making.Aggregate data is generally involved in multi-table joining and aggregating the query data.Enhancing the performance of these operations become the key to improve operational response rate of OLAP.In this paper，an improved aggregation algorithm is proposed based on group number—IMuGA，which takes full advantage of the time dimension table’s particularity，directly gets the group attribute value through the fact table，greatly reduces query times on time dimensional table in multi-table join and improves the efficiency of OLAP query.Experimental results show that the algorithm is effective.

Key words: data warehouse, Online Analytical Processing（OLAP）, aggregate query, multi-table join

摘要： ROLAP是OLAP（联机分析处理）中使用最广泛的一种类型，其主要功能是管理决策所需要的总结数据。总结数据一般都涉及多表连接和分组聚集操作，提高这些操作的性能成为提高OLAP操作响应速度的关键。为此，提出一种基于分组序号的新聚集算法IMuGA。算法充分利用时间维表特殊性，通过对事实表关键字直接获得分组属性值，减少了多表连接中时间维度的连接次数，提高了联机分析处理查询效率。实验结果表明，该算法是有效的。

关键词: 数据仓库, 联机分析处理, 聚集查询, 多表连接