Skip to Main Content
Data transformation problems are very common and are challenging to implement for large and complex datasets. We describe a new approach for specifying data mapping transformations between XML schemas using a combination of automated schema analysis agents and selective user interaction. A graphical tool visualizes parts of the two schemas to be mapped and a variety of agents analyze all or parts of the schema, voting on the likelihood of matching subsets. The user can confirm or reject suggestions, or even allow schema matches to be automatically determined, incrementally building up to a fully-mapped schema. An implementation of the mapping specification can then be generated.