Horizontal Aggregations for Mining Relational Databases
Existing SQL aggregations have limitations to prepare data sets because they return one column per aggregated group using group functions. A significant manual effort using a compliant programming language is required to build data sets, where a horizontal layout is required. Earlier a simple, yet powerful, methods(CASE,PIVOT,SPJ) to generate aggregated columns in a horizontal tabular layout were developed. Both CASE and PIVOT evaluation methods are significantly faster than the SPJ method. The authors propose to use a technique called Generalized Projections (GPs) to improve the performance of SPJ method. The proposed technique pushes down to the lowest levels of a query tree aggregation computation, function computation and duplicate elimination.