No abstract

In this paper we present a technique of compressing bitmap indexes for application in data warehouses. The developed compression technique, called Run-Length Huffman (RLH), is based on the run-length encoding and on the Huffman encoding. RLH was implemented and experimentally compared to the well known Word Aligned Hybrid bitmap compression technique that has been reported to provide the shortest query execution time. The experiments discussed in this paper show that RLH offers shorter query response times than WAH, for certain cardinalities of indexed attributes. Moreover, bitmaps compressed with RLH are smaller than corresponding bitmaps compressed with WAH. Additionally, we propose a modified RLH, called RLH-1024, which is designed to better support bitmap updates.

No abstract