Research of key technology about column store to row store in multi-source heterogeneous database

Abstract

Although at present a lot of big data use the ways of column store, traditional row store is still the mainstream storage way of relational database management system. There is no universal transformation tool facing of the requirements which column stored be transform into row store in heterogeneous database integration systems. The transpose mapping table of column store to row store, data extraction process based on this mapping table, corresponding transposing algorithm are researched and the effectiveness of these key technologies are verified through examples and implementation in this paper. This result is suitable to data extraction from column store to row store for different source table to destination table. Any prerequisites need not be set to the table structure and the type of database of source table and destination table in this research. Therefore, this result possess good generality and compatibility to heterogeneous data sources.