|
|
Constraints and optimization
Executive Summary
We study optimization methods motivated by logical redundancy and
physical independence in mediator-like components.
We have identified
classes of queries and constraints over complex values with
dictionaries, generalizing the relational conjunctive queries and
embedded dependencies, as well as many interesting examples of complex
value and oodb queries and integrity constraints for which we can show
that some important classical results on containment, dependency
implication, and chasing extend and generalize.
Using high-level
representations with the data structure of dictionaries (finite
functions) we are able to represent the physical access structures as
views over the integrated logical schema of the sources.
Dictionaries reflect directly the efficiency of the representation and
facilitate cost estimation.
We define and enumerate in a novel manner a search space for query
plans that incorporate the physical access structures as views on the
logical schema.
We capture the views through constraints (dependencies) and we use our
previous work on the ``equational'' chase to rewrite with these
constraints.
This makes our method easy to integrate in the important
rule-based optimization paradigm, and easy to combine both with
conventional techniques and with our previous approach to semantic
optimization.
Project Members
Val Tannen Alin Deutsch Lucian Popa Arnaud Sahuguet
Publications
|