Incremental Loading of Object Databases

Abstract

Object-oriented and object-relational databases (OODB) need to be able to load the vast quantities of data that OODB users bring to them. Loading OODB data is significantly more complicated than loading relational data due to the presence of relationships, or references, in the data. In our previous work, we presented algorithms for loading new objects that only share relationships with other new objects. However, it is frequently the case that new objects need to share relationships with objects already in database. In this paper we propose using queries within the load data file to identify the existing objects and suggest using parameterized functions to designate similar queries. We then propose a novel evaluation strategy for the queries that defers evaluation until all the queries can be evaluated together. All of the instantiations of a single query function can then be treated as a join between the parameters and the collection over which the function ranges, rather than evaluated as individual queries. We implement both traditional one-query-at-a-time evaluation strategies and our new strategy in a load algorithm for the Shore persistent object repository and present a performance study showing that the new strategy is at least an order of magnitude better when there are many relationships to objects already in the database