An indexing techniques for object-oriented databases

The basic characteristics of object-oriented data models and query languages are summarized. An indexing technique is presented that supports an efficient evaluation of nested predicates for queries involving class hierarchies. A preliminary comparison of the proposed indexing technique with other techniques is presented. A first extension of this work is to define cost formulas for the proposed index organization and to compare it with other organizations, such as the multi-index and inherited-multi-index organizations. A second extension is to define indexing techniques to support alternative predicates on properties that are semantically equivalent. In addition to these, an important issue concerns indexing support for the use of methods in queries.<<ETX>>