Recently, multidimensional data is produced in various domains; because a large volume of this data is often used in complex analytical tasks, it must be stored compactly and able to respond quickly to queries. Existing compression schemes well reduce the data storage; however, they might increase overall computational costs while performing queries. Effectively querying compressed data requires a compression scheme carefully designed for the tasks. This study presents a novel compression scheme, SEACOW, for storing and querying multidimensional array data. The scheme is based on wavelet transform and utilizes a hierarchical relationship between sub-arrays in the transformed data to compress the array. A result of the compression embeds a synopsis, improving query processing performance while acting as an index. To perform experiments, we implemented an array database, SEACOW storage, and evaluated query processing performance on real data sets. Our experiments show that 1) SEACOW provides a high compression ratio comparable to existing compression schemes and 2) the synopsis improves analytical query processing performance.
Download