Schema versioning and database conversion for bi-temporal databases

Schema evolution and schema versioning are two techniques used for managing database evolution. Schema evolution keeps only the current version of a schema and database after applying schema changes. Schema versioning creates new schema versions and converts the corresponding data while preserving the old schema versions and data. To provide the most generality, bi-temporal databases can be used to realize schema versioning, since they allow both retroactive and proactive updates to the schema and database.; This dissertation first overviews the primitive issues of schema evolution and database conversion in a general database aspect. We then in detail explore and formalize the existing research of schema versioning and database conversion techniques for temporal databases: Single Table Versioning (STV) and Multiple Table Versioning (MTV).; We also list the problems of the STV and MTV approaches. We then propose the Partial Table Versioning approach (PMTV) to solve those problems. Comparisons are made among these three approaches in terms of storage space and search time for the current object version. The result shows the proposed approach largely reduce the complexity.; The problem of the Partial Table Versioning approach is that it requires TemporalNatural Join and TemporalOuter Join to retrieve a complete object information, and both of the operations are very expensive. We propose the Bi-temporal EntityJoin method associated with a bi-temporal index structure to reduce the overhead. We also do the analysis and performance comparison. The result shows it performs better than other approaches adopting only bi-temporal index.