The basics of data normalization

In 1NF, you divide the base data into logical units called entities or tables. When you design each entity or table, you assign the primary key to it, which uniquely identifies each record inside the table. You create a separate table for each set of related attributes. There can be only one value for each attribute or column heading. The 1NF eliminates the repetition of groups by putting each one in a separate table and connecting them with a one-to-many relationship.

The objective of 2NF is to avoid the duplication of data between tables. In 2NF, you take data that is partly dependent on the primary key and enter it into another table. The entity is in 2NF when it meets all of the requirements of 1NF and has no composite primary key. In 2NF, you cannot subdivide the primary key into separate logical entities. You can, however, eliminate functional dependencies on partial keys by putting those fields in a separate table from the ones that are dependent on the whole key.

The 3NF objective is used to remove the data in a table that is not dependant on the primary key. In 3NF, no non-key column can depend on another non-key column, so all of the data applies specifically to the table entity. The entity is in 3NF when it meets all of the requirements of 1NF and 2NF and there is no transitive functional dependency.