Database normalisation by D.Lukachuk
- 1. Database Normalization Principals of Database Normalization by Dmytro Lukachuk
- 2. Overview • Normalization, main ideas; • Normal forms; • Normalization and Denormalization; • Practical examples.
- 3. Normalization • Normalization is the process of organizing the attributes and tables to minimize redundancy. Wikipedia
- 4. Aims of normalization: • Minimize logical redundancy; • Free the DB of modification anomalies; • Minimize redesign when extending DB structure; • Simplification for applying integrity constraints. Edgar Frank Codd
- 5. Other suggestions: • The aim of Normalization is not to decrease or increase performance; • The aim of Normalization is not to decrease or increase physical memory size of DB.
- 6. So what is Normalization? • The process of transformation of database relations to the structure that fits normal form (NF).
- 7. Normal Form • A set of requirements (rules, principals) that satisfy relations in DB.
- 8. Normal forms: • 1st Normal Form (1NF); • 2nd Normal Form (1NF); • 3rd Normal Form (3NF); • Boyce-Codd Normal Form (BCNF); • 4th Normal Form (4NF); • 5th Normal Form (5NF); • Domain/key Normal Form (DKNF); • 6th Normal Form (6NF).
- 9. 1st Normal Form Principals • No top-to-bottom ordering to the rows; • No left-to-right ordering to the columns; • No duplicate rows; • Every cell contains only one value; • All columns are regular (no hidden components).
- 10. 2nd Normal Form Principals • Relations are in 1st Normal Form; • Every non-prime attribute is dependent on candidate key.
- 11. 3rd Normal Form • Relations are in 2nd Normal Form; • Non-prime attribute can be determined only by the candidate key and not any other non- prime attribute (transitive dependency).
- 12. Employee -> Department Department -> Telephone Employee -> Telephone (transitive dependency)
- 13. Boyce-Codd Normal Form • Relations are in 3rd Normal Form • Primary key does not depend on non-prime attribute. Relation is in 3NF but not in BCNF when relation has two or more candidate key that are composite and have common attribute
- 14. Denormalization • Transformation of DB structure to the form that does not satisfy normal forms. • Process of optimizing performance by adding redundant data or by grouping data.
- 15. Advantages of Normalization • Smaller • Fast updates • Fast inserts • Less need for heavy group by or distinct queries as data is not duplicated
- 16. Disadvantages of Normalization • Joins • Bad indexing strategy
- 17. Advantages of Denormalization • No joins • Fast select • Efficient index usage
- 18. Disadvantages of Denormalization • Costly inserts • Costly updates
- 19. Compare Normalization: SELECT product_name, order_date FROM orders INNER JOIN products USING(product_id) WHERE product_name like 'A%' ORDER by order_date DESC Denormalizaion: SELECT product_name, order_date FROM orders WHERE product_name like 'A%' ORDER by order_date DESC
- 20. Slider structure analysis: Slider table
- 23. Slide table
- 26. What is the best way? The best way is to – MIX principals of normalization with denormalization approach depending on certain application purposes.