Defining individual fields
For each field, you define its name, type, size, decimals (if a numeric or float field), and index (optional). The specifics of field types are discussed in the following section. Here are some overall guidelines:
Use one piece of information per field. For example, put city, state, zip code, and country data into separate fields, because you might want to process the information in each field separately. However, do not split certain information, such as street number and street name, unless you need to process rows by street name or street number separately.
Keep field sizes to a minimum, without being excessively restrictive, to conserve disk space. If you intend to total a numeric field, you must define a field large enough to hold the total, not just individual values.
For indexed fields, use abbreviated codes instead of long character fields wherever possible. For example, instead of duplicating the entire customer name in every order, use a short customer code to simplify data entry, indexing, and linking. This results in more efficient indexes and makes it easier to update information.
Define fields in a logical order in the table. The order you define is the default way in which users will see the table. In general, put indexed fields toward the beginning of the table, and put similar information together in a sensible sequence.
Use descriptive, unique field names. Be consistent when naming fields that contain similar data. Standardize field names shared across tables if possible (this is not permitted with some SQL databases).