Generalized Projections: A Powerful Approach To Aggregation
Abhishek Gupta, Venky Harinarayan, Dallan Quass · 1995
In this paper we introduce generalized projections (GPs), an extension of duplicate-eliminating projections, that capture aggregations, groupbys, conventional projection with duplicate elimination (distinct), and duplicate-preserving projections in a common unified framework. Using GPs we extend well known and simple algorithms for SQL queries that use distinct projections to derive algorithms for queries using aggregations like sum-max-min-count and avg. We develop powerful query rewrite rules for aggregate queries that unify and extend rewrite rules previously known in the literature. We then illustrate the power of our approach by solving a very practical and important problem in data warehousing: how to answer an aggregate query about base tables using materialized aggregate views (summary tables). Keywords: aggregation, data warehousing, materialized views, query optimization 1 Introduction With the growing number of large data warehouses for decision support applications, effi...