Optimizing Queries Using Materialized Views: A Practical, Scalable Solution
Jonathan Goldstein (Microsoft) and Per-Åke Larson (Microsoft)
This paper from the SIGMOD 2001 Conference describes a new technique for rewriting queries using materialized views. The paper introduces an efficient view matching algorithm and a new index structure, the filter tree, on view definitions that is used to determine which parts of a query can be computed from materialized views. The technique incurs very little penalty in the time taken to generate the query plan. The paper has been highly influential in subsequent research on the use of materialized views for query processing and in other contexts such as federated systems, probabilistic databases and XML databases; and it has influenced the design of query optimizers in at least two commercial database systems.