ABSTRACT
The third normal form logical data model is used as the basis for creating a physical database, as it represents a clean representation of the user’s data requirements where each non-key attribute appears only once, where it belongs. A physical data model is created by implementing a table for each entity in the logical model, with attributes mapping to columns in their respective tables. However, in practice, this implementation generally results in poor performance for many critical user processes. To produce a physical solution with satisfactory performance levels, the design team analyzes data access paths for high performance, frequently used processes, then makes changes in the physical model that results in meeting data access requirements with fewer I/Os. This normally introduces redundant information which the design team manages through their application software.
In creating the physical database, other key decisions must be made. What data format is appropriate for each column? Can the key column(s) for each table be efficiently implemented? What indexes are needed? How are Referential Integrity constraints implemented? Starting with the third normal form logical data model from Chapter 2, the current chapter addresses these questions in detail, deriving a physical database using both Microsoft Access and Structured Query Language Server.
