next up previous contents
Next: Contents

User-Defined Aggregates in Database Languages

This work was done by Punit Bhargava as Master's thesis under the guidance of Dr. Carlo Zaniolo. The postcript version of the thesis can be downloaded from here . Any comments or suggestions could be mailed to punit@cs.ucla.edu.

Abstract:

Applications, such as OLAP, data warehousing, and data mining, are steering today's database industry. These applications process highly complex queries and make extensive use of complex aggregates such as online aggregation and data cubing. Current SQL2 compliant database systems cannot support these complex aggregates. In fact, SQL2 only supports five aggregates and provides no mechanism for defining new aggregates. Constructs for specifying new aggregates (i.e., user defined aggregates) are part of the forthcoming SQL3 standards. In this thesis, we propose SQL-AG, a language and a system to support user-defined aggregate in SQL. The proposal extends SQL3 user-defined aggregates and removes some of their limitations to allow the definition of on-line aggregates, data cubes and other aggregates used in data mining. This report describes the design and implementation of the SQL-AG system. We begin with an introduction to the problem, followed by an analysis of the design issues related to SQL-AG. All details of the design and implementation of SQL-AG, along with a discussion of the implementation environment based on Oracle Corporation's server and PL/SQL, are presented.





Punit Bhargava
Wed Mar 11 18:50:53 PST 1998

All contents copyright © 1998, Punit Bhargava.