A Solution to Ragged Dimension Problem in OLAP

摘要

One problem facing many existing OLAP (On-Line Analytical Process) systems is the so-called ragged dimensions. Ragged dimensions occur if the logical parents of some members in a dimension hierarchy are two or more levels apart. In other words, there exist empty holes in the dimension hierarchy. The problems caused by ragged dimension are two-fold. First, aggregation of measure data could be incorrect. Second, the pre-computation strategy, the most prevalent technique used to speed up query processing in current OLAP system, could be rendered invalid. This paper proposes a simple yet efficient solution to remedy the ragged dimension problem for existing OLAP systems.