Adaptively incremental dictionary compression method for column-oriented database

摘要

As an efficient method, the dictionary compression approach has attracted much interest and been widely used in column-oriented database systems. The core idea is to represent a longer string by a shorter integer, which significantly reduces the size of data and then improves the input-output (I/O) performance. To further reduce the space overhead of the dictionary compression method based on static increment, a novel adaptively incremental dictionary compression encoding method is developed in this paper. The developed method makes full use of the statistical characteristics of the data to reduce the footprint of the dictionary in the memory, which is also shown more flexible in reality. Furthermore, a better performance of query execution can be obtained by the adoption of such compression method in the query plan. The simulation results under different experiments show the effectiveness and efficiency of our proposed method.