What defines third normal form (3NF) in database normalization?

Third normal form (3NF) is achieved when a table is already in second normal form (2NF) and additionally has no transitive dependencies. This means that all non-key attributes are fully functionally dependent on the primary key and do not depend on other non-key attributes. In essence, 3NF aims to eliminate redundancy and ensure that data integrity is maintained, allowing for efficient data management.

To be in 3NF, a database table should not have any attribute that depends on another non-key attribute. For instance, if you have a table storing information about students where one column lists the student's major and another column references the department head of that major, you would have a transitive dependency because the department head is dependent on the major, which in turn is dependent on the student ID (the primary key).

In contrast, the other options do not correctly define 3NF. For example, the requirement of having three primary keys or at least five attributes is not relevant to normalization forms, as normalization pertains to relationships and dependencies among attributes rather than the quantity of keys or attributes. Additionally, allowing duplicate records contradicts the purpose of normalization, which is to reduce redundancy and ensure that each data point is stored uniquely.