Blackwell Reference Online

Extract

Given a description of the logical data content and activities for an application, physical data modeling is concerned with designing efficient structures and algorithms for the storage and retrieval of those data. A physical data model describes the details of how data are stored and accessed. The efficiency of a physical data model is determined by such factors as: the volume and intrinsic structure of the data, the frequency and content of retrievals and updates, and the accessing characteristics of the data storage media. Figure 4  Data model derived from invoice document in figure 3 There are two major aspects of a physical data model: record structures and access paths . Record structures define the representation, grouping, and interconnection of elementary data units (attributes and instances). Access paths define the algorithms and system data used to store and retrieve the data. Given the characteristics of common data storage media, a general rule for physical data modeling is to store data together that are used together. A database management system (DBMS) can be characterized by the record structures that it supports. Relational database management systems typically restrict record structures to flat tables interconnected via data values (foreign keys). Object‐oriented database management systems typically support complex record structures in which objects of ... log in or subscribe to read full text