Data Warehouse Native Feature Based OLAP Querying with Keywords
XIAO Min,CHEN Ling,XIA Hai-yuan,CHEN Gen-cai
DOI: https://doi.org/10.3785/j.issn.1008-973x.2012.06.003
2012-01-01
Abstract:In order to support naive users to use on-line analytical processing(OLAP) tools,this work proposed a native feature of the dimension and its attributes for the multidimensional model of data warehouse based OLAP query method,which combined OLAP and search engine-alike methods.The method firstly created column based full-text index for dimension tables,secondly generated hit groups in terms of the keywords provided by the user,then constructed a candidate result set by joining these hit groups,and finally presented the ranked results to the user.Based on the features that a user cared much more about the aggregated data and the imbalance of the dimensions and their attributes,irrelative attributes and repeated column values were filtered in proposed method to relieve the negative effects on the ranking results.A weighted coefficient,named dimensional level coefficient(DLC) was also introduced to the text ranking arithmetic.Experiments were conducted on the FoodMart and AdventureWorks provided by Microsoft SQL Server to confirm how these factors influenced the hitting rate.The results indicated that the proposed method achieved a higher hitting rate than the keyword-driven analytical processing(KDAP) methods for the first candidate.