Transforming Heterogeneous Data with Database Middleware: Beyond Integration.

摘要: Many applications today need information from diverse data sources, in which related may be represented quite differently. In one common scenario, a DBA wants to add new source an existing warehouse. The the not match warehouse schema. also partially redundant with that warehouse, or formatted Other integrate more dynamically, response user queries. Even using single often want present it form other than is stored in. For example, publish some particular XML DTD, though form. each of these scenarios, sets must mapped into target representation. Needed transformations include schema (changing structure data) [BLN86, RR98] and transformation cleansing format vocabulary eliminating at least reducing duplicates errors) [Val, ETI, ME97, HS95]. area, there broad range possible transformations, simple complex. Schema have typically been studied separately. We believe they handled together via uniform mechanism. Database middleware systems [PGMW95, TRV96, ACPS96, Bon95] multiple sources. To effective, such provide integrated schemas, able transform different sources answer queries against power their query engines ability connect several makes them natural base for doing complex as well. this paper, we look database tranformation engines, discuss when how transformed users need.