MULTIDIMENSIONAL OLAP QUERY PROCESSING METHOD FOR COLUMN-ORIENTED DATA WAREHOUSE

摘要

Disclosed is a multidimensional OLAP query processing method for column-oriented data warehouses. The method decomposes an OLAP query into a bitmap filtering operation, a grouping operation, and an aggregation operation. In the bitmap filtering operation, implementing the predicate in the dimension table and generating a predicate vector bitmap, and converting the connection operation into the access operation to the direct dimension table record through a proxy key address mapping, thereby enabling access by position; in the grouping operation, pre-generating a group unit from a fact table record meeting the filtering condition according to the grouping attribute in the SQL command, and assigning the incremental ID; in the aggregation operation, performing one column scan on the measurement attribute of the fact table and enabling grouping and aggregation calculation according to the grouping items of the filtering and grouping vector in the fact table. The present invention only performs one column scan on the fact table to complete all OLAP processing tasks, thereby avoiding the cost of multiple scanning.