Data consistent issue

• repetition of data는 database의 사이즈가 커지는 데 문제점이 있지만...

  Insertion anomaly

  • inserting redundant data for every new record

  Deletion anomaly

  • Loss of related data, when some data is deleted

  Update anomaly

  • When updating certain information, every single record must be updated

Decomposition

• Normalization
  • decomposition of relational schemas
  • only directly related data composes a relation
    • 직접적으로 관련이 있는 데이터만 테이블에 넣음.
    • 하나의 테이블을 여러 개로 나눔.

Less redundancy

• Uses smaller disk storage
  • insertion, deletion, and update anomalies
• 모든 분해가 좋은 건 아니다.

Lossy decomposition

• a decomposition from which the original relation cannot be reconstructed
  • data components간의 연결 정보를 손실한다.
  • 더 많은 tuple이 생길 가능성이 있다.

Lossless Decomposition

• lossless decomposition이 되어야 한다.

Normalization

• Database normalization

  • reduce data redundancy
  • improve data integrity
    • a series of normal forms

  process

  • decompose relations to suppress data anomlies
  • make sure the decomposition is lossless

https://yaboong.github.io/database/2018/03/09/database-normalization-1/

• 대체