Mapping Adaptation under Evolving Schemas Yannis Velegrakis, Renée J. Miller, and Lucian Popa   View Paper (PDF)   Return to Managing derived Data (Session B6) Abstract To achieve interoperability, modern information systems and e-commerce applications use mappings to translate data from one representation to another. In dynamic en- vironments like the Web, data sources may change not only their data but also their schemas, their semantics, and their query capabilities. Such changes must be reflected in the mappings. Mappings left inconsistent by a schema change have to be detected and updated. As large, com- plicated schemas become more prevalent, and as data is reused in more applications, manually maintaining map- pings (even simple mappings like view definitions) is be- coming impractical. We present a novel framework and a tool (ToMAS) for automatically adapting mappings as schemas evolve. Our approach considers not only local changes to a schema, but also changes that may affect and transform many components of a schema. We con- sider a comprehensive class of mappings for relational and XML schemas with choice types and (nested) constraints. Our algorithm detects mappings affected by a structural or constraint change and generates all the rewritings that are consistent with the semantics of the mapped schemas. Our approach explicitly models mapping choices made by a user and maintains these choices, whenever possible, as the schemas and mappings evolve. We describe an imple- mentation of a mapping management and adaptation tool based on these ideas and compare it with a mapping gen- eration tool. ©2004 Association for Computing Machinery