Data analysis | Database

Normalisation - relational data analysis

• Normalisation:
  • Structuring data to achieve data integrity.
  • The process of converting complex data structures into simple stable data structures.
  • It is an approach that examines data and groups data items in a way that seeks to minimise the impact of change by reducing the duplication of data, represents data in form of a two dimensional table (refer Relational Database)
• Objectives:
  • Prevent insert/update and delete anomolies
  • Minimise redundancy (produces controlled redundancy)
• An entity is normalised when it is
  " Dependent on"
  • the key                     = first normal form
  • the whole key          = second normal form
  • nothing but the key = third normal form

Once you have normalised your data you may have to go back and review your Entity Relationship diagrams

Notation

• Entity = A related set of data eg. Supplier.
• Data element = A unit of data forming an item or a field eg. SuppName
• KEY unique identification key
  • Atomic key = key with single attribute
  • Compound key = key with mutiple attributes
  • Primary key = A unique identifier by which the data may be located eg. SupplierId.
  • Foreign key* = key of a related table
• ATTRIBUTE
  • Mandatory = must have (not null).
  • Optional =  could have null or missing value

• Projection is the elimination of columns on a table, removing duplicate rows
• Join is the inverse of projection