Relational Model

The relational model, introduced by Edgar Codd in 1970, is the theoretical foundation for relational DBMS. It represents data as relations (tables), where each relation is a set of tuples (rows) with attributes (columns). All data manipulation is performed through Relational Algebra operations.

Core Terminology

Formal Term	Common Term	Description
Relation	Table	A set of tuples with a defined schema
Tuple	Row / Record	A single data entry in a relation
Attribute	Column / Field	A named property with a domain (type)
Domain	Data type	The set of allowed values for an attribute
Schema	Table definition	The structure (attribute names and domains)
Degree	Column count	Number of attributes in a relation
Cardinality	Row count	Number of tuples in a relation

Integrity Rules

Entity Integrity

Every relation must have a Primary Key, and no primary key attribute can be NULL.

Referential Integrity

Every Foreign Key value must match an existing Primary Key value in the referenced relation, or be NULL.

Domain Integrity

All attribute values must come from the attribute’s defined domain (enforced via SQL Data Types and CHECK constraints).

Properties of Relations

• No duplicate tuples — Guaranteed by the Primary Key
• Tuples are unordered — There is no inherent row order (ORDER BY is a query operation)
• Attributes are unordered — Column order doesn’t matter
• Attribute values are atomic — Each cell contains a single value (1NF)

Schema vs Instance

• Schema — The structure definition (table name, column names, types, constraints). Rarely changes.
• Instance — The current set of data (rows) in the table. Changes with every INSERT/UPDATE/DELETE.

From Conceptual to Relational

The Entity-Relationship Diagram (ERD) captures the conceptual model, which is then mapped to the relational model through Normalization:

• Entities → Tables
• Attributes → Columns
• Relationships → Foreign keys and junction tables
• Constraints → PRIMARY KEY, FOREIGN KEY, UNIQUE, CHECK