Per the IBM Dictionary of Computing, a data dictionary is a "centralized repository of information about data such as meaning, relationships to other data, origin, usage, and format." Essentially, it is documentation or data about data in a file or database, also called metadata.
Typical Items In A Data Dictionary
For each field or column in a database table, describe the following elements:
Item
Contents
File/Table Name
The name of the file or table where this field/column is stored (origin)
File/Table Description
A description of the purpose/usage of the file/table.
Field/Column Name
Data Type
Character (Char), numeric (N) including number of decimals, integer (Smallint, Int), memo, binary, varchar, etc.
Position
The starting position of field in record layout (if not in a database table)
Length
Field length
Relationships
If a database table, describe whether column is a Primary (PK) or Secondary Key (SK) to x column in x table.
Meaning
A detailed description of what the data means, including explanations of codes, etc. For example: M = Male, F = Female.
Usage
A detailed description of how the data is used in the system including special processing considerations. For example: “Records with code x must be filtered out to balance to the general ledger total.”
Required/Null
An indication as to whether this field is allowed to have no data (null).
Remember, the meaning and usage of a particular data element may not be exactly what is provided in the system documentation. The users of the system may be storing other data in that field/column for an entirely different purpose. So, be sure to confirm the data meaning and usage with an expert user of the system!
Example Database Data Dictionary
Below is an example of a typical data dictionary obtained from the documentation of a database. This is a good start, but should be enhanced with usage details where applicable to be more useful to end