SQL — Normalized Data

Normalized data is defined as data in individual tables must contain atomic, semantically related sets.

Atomic means that each column in a table must not be able to be broken down in to sub-components. For example, a part number should not have the first two bytes contain a special meaning and the next four bytes contain a different special meaning. In this situation, it is more appropriate to describe this data as two columns instead of one.

Semantically related sets of each file should contain information that relates directly to the primary key. This implies that each record in the file will exhibit the exact same format. Therefore, files with multiple record types are non-normalized.

The BASIS DBMS provides the tools to address non-normalized data found in many existing legacy Business BASIC applications. The BASIS DBMS provides a relational view mechanism which allows the logical separation of semantically related information from a base file and present it as its own table.