Mapeamento de esquemas conceituais geográficos para esquemas GML e esquemas físicos de bancos de dados espaciais

Abstract

Geographic conceptual modeling, along with traditional conceptual modeling, is fun-damentally important in the design of geographic applications. Geographic conceptual models provide primitives to represent the geometry and the topology of geographic data, which are generally stored in GML documents or spatial databases. GML isalso widely used nowadays in data exchange among geographic applications or through the Web. Geographic databases provide support for spatial data handling, including geometric and topologic functions, but do not facilitate the implementation of spatial integrity constraints. This dissertation defines the mapping of OMT-G geographic conceptual schemas into GML schemas and physical geographic database schemas. For that purpose, rules and mapping algorithms are proposed in order to preserve the structures and semantics of conceptual schema, such as spatial and non-spatial classes and relationships. GML schemas are generated using a strategy that is able to reduce the use of references and increase the nesting of elements. Reducing references is important because it leads to faster validation of GML documents, while nesting improves the efficiency of queries over GML documents. A comparison between the proposed method and related work is presented, assessing the mapping techniques and query processing times over queries on GML documents. Results show that our mapping method was able to generate time-efficient results. Physical schemas are generated considering both the OMT-G class diagram primitives and spatial integrity constraints that can be deduced from analyzing the conceptual diagrams. The resulting schemas include tables withgeometry columns, spatial indexes, spatial-consistency-checking functions, and other elements.