Organizing a database

The following describes the process of data normalization. Normalization is a method of logical data organization that minimizes data redundancy and maximizes data independence.

Data normalization is commonly used in the design of relational databases. Although a TPFDF database is hierarchical, it can still benefit from data normalization in the first stage of its design. In later stages, you will need to move away from rigid normalization to achieve the performance standards necessary for a high-volume transaction processing system. These stages are described in Optimizing the Database Design and Mapping Tables to TPFDF Files.

Normalization reduces the data to a minimal form, providing a clearer path for subsequent stages in the design process. The normalization process is always preceded by an analysis of the data to be stored. For now, it is assumed that you have already determined the following:

• The entities for which data is stored.
• The attributes (data fields) that will be stored for each entity.

In the following, you can assume that each table will map to a physical TPFDF file, and each row in a table to a TPFDF logical record (LREC).

TPFDF databases have the following characteristics:

• Many users
• A large number of files
• Different means of accessing files
• A high transaction rate.

Consider these characteristics as you work through the design of your TPFDF database.