Fixed-length string compression for direct operations in column-oriented databases
Ke Yan, Meiyi Xie, Hong Zhu · 2013
Compression is one of the most important techniques in column-oriented database systems development. For fixed-length string typed columns, both heavyweight and lightweight compression schemes have limitations. In this paper, we propose a compression scheme, called FSC (Fixed-length String Compression), to achieve good compression ratio and support direct queries on compressed data without decompression in advance. The main idea of FSC is to vertically partition a fixed-length string typed column into sub-columns, which are compressed by different lightweight compression methods. Moreover, we present a search method, which are called FSC-search, to search on compressed data directly. Intensive experiments show that FSC not only achieves good compression ratio, but also improve query performance by supporting direct searching on compressed data.