ITPEC IP April 2023 Question 27

To retain the consistency of data — normalisation removes redundant data so that each fact is stored in one place only.

When the same fact is stored in many rows, for example a customer's address repeated in every order row, changing it in some rows and not others leaves contradictory data (an update anomaly); insertion and deletion anomalies arise the same way. Normalisation splits tables by their functional dependencies (first, second and third normal form) so that each fact is held once and the tables are linked by keys. A change is then made in one place, and the data stays consistent.

Answer (b)

Why not others:

• •(a) — normalisation is the logical design of tables; compressing data to save space is a separate technique
• •(c) — leakage is prevented by access control and encryption, not by the table structure
• •(d) — simultaneous access by many users is handled by the DBMS's exclusive control (locking) and transactions

Key rule: Normalise to store each fact once and avoid update anomalies; the cost is more joins when reading, which is why data warehouses are sometimes deliberately denormalised.

Databases · AI-written, may contain errors