Short Cuts for Traversals in Object-Oriented Database Systems
W.C. Lee, Dik Lun Lee · 1994
We introduce a new class of indexing organizations for nested query processing in object-oriented database systems. Our methods maintain direct links connecting objects in two separate classes for fast object traversals. We develop cost models for di erent implementations of the indexing organizations and compare their performance. The clustering index has been shown to be the best in terms of storage and retrieval costs. However, the non-clustering and semi-clustering indexes are also very e cient when compared to traditional object traversal mechanisms. 1