A PostGIS extension to support advanced spatial data types and integrity constraints

Abstract

Geometric primitives defined by OGC and ISO standards, implemented in most modern spatially-enabled database management systems (DBMS), are unable to capture the semantics of richer representation types, as found in current geographic data models. Moreover, relational DBMSs do not directly extend referential integrity mechanisms to cover spatial relationships and to support spatial integrity constraints. Rather, they usually assume that all spatial integrity checking will be carried out by the application, during the data entry process. This is not practical if the DBMS supports many applications, and can lead to redundant and inconsistent work. This paper presents AST-PostGIS, an extension for PostgreSQL/PostGIS that incorporates advanced spatial data types and implements spatial integrity constraints. The extension reduces the distance between the conceptual and the physical designs of spatial databases, by providing richer representations for geo-object and geo-field geometries. It also offers procedures to assert the consistency of spatial relationships during data updates. Such procedures can also be used before enforcing spatial integrity constraints for the first time. We illustrate the use of AST-PostGIS on an urban geographic database design problem, mapping its conceptual schema to the physical implementation in extended SQL.