Multitemporal Conditional Schema Evolution

Jensen, OG; Bohlen, MH
Jensen, O
Bohlen, M
Proc. 3rd ER-Workshop Evolution and Change in Data Management (ECDM 2004), LNCS 3289
Citations range: 
1 - 9
Jensen2004MultitemporalConditionalSchemaEvolution.pdf237.57 KB

Schema evolution is the ability of the database to respond to changes in the real world by allowing the schema to evolve. The multidimensional conditionally evolving schema(MD-CES) is a conceptual model for conditional schema changes, which modify the schema of those tuples that satisfy the change condition. The MD-CES is lossless and preserves schemas, but has an exponential space complexity. In this paper we restrict conditional schema changes to timestamp attributes. Specifically, we develop 1D-CES for schema versioning over one time dimension, and 2D-CES for schema versioning over two time dimensions. We show that the space complexity of these new evolution models is linear or polynomial. 1D-CES and 2D-CES are compared to temporal schema versioning, and we show that, unlike valid time versioning, they are lossless and achieve the same space complexity as temporal versioning if the schema changes are ordered.