In Relational Databases, any field that could be a component of a Candidate Key is called a "Prime Attribute". If a field could not be a part of a candidate key, that's considered a "non-prime attribute".

Illustrative Example:

Given a table had the following columns:

• person ID
• first name
• last name
• phone number
• email
• height
• weight
• age
• gender

You could feasibly see 4 sets of Candidate Keys like the following:

• A - person ID
  • an example of Surrogate Keys, by convention forced to be unique (if otherwise meaningless)
• B - phone number
  • usually unique to the individual, but sometimes shared
• C - email address
  • usually unique to the individual, but sometimes shared
• D - first name + last name
  • naming collisions happen in large enough datasets

Thus, the Prime and non-Prime attributes are as follows:

• PRIME - person ID
• PRIME - first name
• PRIME - last name
• PRIME - phone number
• PRIME - email
• NON-PRIME - height
• NON-PRIME - weight
• NON-PRIME - age
• NON-PRIME - gender

Source

https://en.wikipedia.org/wiki/Candidate_key

Related Notes

Relational Databases

Candidate Key