Query Processing Techniques for Arrays

Arunprasad P. Marathe       Kenneth Salem*
University of Waterloo       University of Waterloo
apmarath@db.uwaterloo.ca       kmsalem@uwaterloo.ca

Abstract

Arrays are an appropriate data model for images, gridded output from computational models, and other types of data. This paper describes an approach to array query processing. Queries are expressed in AML, a logical algebra that is easily extended with user-defined functions to support a wide variety of array operations. For example, compression, filtering, and algebraic operations on images can be described. We show how AML expressions involving such operations can be treated declaratively and subjected to useful rewrite optimizations. We also describe a plan generator that produces efficient iterator-based plans from rewritten AML expressions. Because the physical operators are iterators, the resulting plans are suitable for pipelined evaluation, allowing for inter-operator parallelism.