
考試只會考1NF、2NF、3NF
標準化越後面切越多,造成join多、效率低
Normalization
- 評估、修正table結構減少data redundancies
- 減少data anomalies
- 由determination來assign attributes
- 若是3NF設計好的話,就會達到4NF。標準化越好時,join就會越多次
- 切越多、越細其實就是資料越分散,導致查詢(join)困難
- Denormalization:Produces a lower normal form
- Results in increased performance and greater data redundancy
Need for Normalization
- designing a new database structure
- Analyzes the relationship among the attributes within each entity
- Determines if the structure can be improved
- Improves the existing data structure and creates an appropriate database design

Table form ⇒ Table name: RPT_FORMAT

造成問題:
- PROJ_NUM為PK的話,會有null,所以用PROJ_NUM+EMP_NUM
- 人為輸入時導致資料不一致(data inconsistencies),如“Elect.Eng.” = “EE”
- 更新資料時產生Update anomalies,如更新EMP_NUM = 105,很多地方都要一起變
- 產生Insertion anomalies,當加入新員工時,沒有正在執行的計畫時,就只能產生一個空的計畫放入新員工