PLR Partitions: A Conceptual Model of Maps

The traditional spatial data models model space in terms of points, lines, and regions. These models employ a disjoint dimension model in which a spatial object can only consist of a zero, one, or twodimensional point set. However, such models cannot adequately represent spatial reality. For example, consider a river network that contains both rivers and lakes. Intuitively, this river network is a single object made up of one-dimensional components (the river segments), and twodimensional components (the lakes). Typically, collection types are used to represent such an object, but they present new problems. In this paper, we propose the PLR partition model that is able to model space in the form of a map geometry that can contain point, line, and region features within the same object. This model solves the problems associated with the traditional spatial data models.