Database Normalization

Five 'rules' to make data normal: 1NF,2NF,...,5NF (NF=normal form)
- each rule builds on another, starting from 1NF
- 1NF/2NF/3NF -> Core Basics (normalization usually means 3NF!)
- 4NF/5NF -> Exceptions

*Note: Normalization is about grouping & connecting data the right way!

1NF is about
- atomic values
- unique identifiers

*Term: Imagine a spreadsheet -> [table] or [entity]

1NF rules
- a cell cannot contain more than 1 value
  = if it does, that column needs to be split into multiple columns
- each row (@record) must be unique
  = look for potential primary key
  = usually, we use system-generated (integers are better!)
- each column name must be unique
- there must be no repeating groups (or cells?)
  = it there are, remove and create new table (1NF!)

2NF rule
- all data/column(s) must depend on the primary key
  = if it does not, must be split into it own table (1NF!)
  = use primary key in new table as column value -> foreign key

3NF rule
- primary key must define all non-key column(s)
  = non-key column(s) must not depend on any other key
  = if this is not met, must create new table and use foreign key to link