Modelling

Database: integrated collection of data

Relational Model

Two parts

Constraints

Domain: each value must an atomic value in the domain e.g., integer, string, date, etc
Key and NULL:
- Superkey: uniquely ids a tuple e.g., set of all attrs
- Key: minimal superkey
- Candidate key: a possible key
- Primary key: the chosen candidate key: cannot have NULL values
Ref Integrity: Foreign key
Functional Dependencies
- fd: when one value determines another e.g., sid -> sname, {sid, cid} -> grade
- armstrong axioms:

Help to prevent anomalies (update, delete, insert) by reducing redundancy

e.g.,

bad:
Student_Course(sid, sname, address, cid, cname, grade) 

good:
Student(sid,sname,address) 
Course(cid,cname)
nrolled(sid,cid,grade)

done through process of decomposition: breaking down schema into set of schemas while preserving:

Various degrees of normalisatin defined by normal forms: Normal Forms