A Parallel Hierarchical Aggregation Algorithm in High Dimensional Data Warehouse

OLAP (on-line analytical processing) queries tend to be complex and ad hoc, often requiring computationally expensive operations such as multi-table joins and aggregation. In the high dimensional data warehouse(DW), we full materialized the data cube impossibly. In this paper, we propose a novel aggregation algorithm, PDHEPA (parallel pre-grouping aggregation based on the dimension hierarchical encoding), to vertically partition a high dimensional dataset into a set of disjoint low dimensional datasets called fragment mini-cubes. PDHEPA uses the small dimension hierarchical encoding and their prefix, so that it can drastically reduce the multi-table join operations. As a result, the method we proposed in this paper can greatly reduce the disk I/Os and highly improve the efficiency of OLAP queries. The analytical and experimental results show that the PDHEPA is more efficient than other existed ones.