EER Modeling: Categorization (Union Types) and Aggregation

Introduction: Handling Complex Real-World Relationships

In basic specialization and generalization, a subclass inherits from a single superclass (or multiple superclasses sharing the same keys).

However, real-world business scenarios often have situations that break standard hierarchy rules:

1. Scenario 1: A car owner can be an individual person, a registered company, or a bank. How do you create an OWNER entity when each category has completely different keys and structures?
2. Scenario 2: A doctor treats a patient during a visit, and that specific treatment needs to be billed by an insurance company. How do you relate a relationship to another entity?

To solve these complex architectural challenges, the EER model introduces Categorization (Union Types) and Aggregation.

1. Categorization (Union Types)

What is a Category (Union Type)?

A Category (or Union Type) represents a single subclass that has two or more distinct superclasses where each superclass represents a fundamentally different entity type with different primary keys.

Key Characteristics of Categorization:

• Selective Inheritance: In ordinary generalization, a subclass inherits all attributes from its superclass. In categorization, an instance in the subclass belongs to only ONE of the superclasses at any given time and inherits attributes only from that specific superclass!
  • If Owner 1 is a PERSON, they inherit Aadhar_No and DOB.
  • If Owner 2 is a COMPANY, they inherit GSTIN and Company_Name.
• Total vs. Partial Category:
  • Total Category: Every instance of the superclasses must be an owner (Drawn with a double line).
  • Partial Category: An entity can exist without being an owner (e.g., Not every person owns a vehicle).

2. Aggregation in the ER Model

The Problem: Relationships Between Relationships

In Chen’s standard ER model, a relationship can only connect entities, never another relationship:

• Suppose EMPLOYEE works on a PROJECT. This is a relationship WORKS_ON.
• Now, suppose management needs to assign specialized MACHINERY to that specific employee-project assignment.
• Drawing a line from MACHINERY into the diamond WORKS_ON is strictly forbidden in basic ER modeling!

The Solution: Aggregation

Aggregation is an abstraction concept where a relationship between two entities is treated as a single higher-level abstract entity.

• By enclosing the entire [EMPLOYEE - WORKS_ON - PROJECT] relationship inside a bounding rectangle, it now behaves as an ordinary entity that can legally participate in a new relationship <USES> with MACHINERY.

3. Comprehensive Comparison: ER Model vs. Enhanced ER (EER) Model

💡 DAVV Exam Preparation Tips

5-Mark Question: “What is Aggregation in the ER model? Why is it required? Explain with a diagram.”

• Explain why relationships cannot directly connect to other relationships in standard ER.
• Define aggregation as treating an entity-relationship pair as a single abstract entity.
• Draw the [EMPLOYEE - WORKS_ON - PROJECT] aggregates to <USES> MACHINERY diagram.

10-Mark Question: “Explain Categorization (Union Types) and Aggregation in the Enhanced ER (EER) model. Provide at least 6 points of comparison between basic ER and EER models.”

• Define Categorization and explain why selective inheritance is needed when superclasses have distinct keys (Person vs Company as Vehicle Owner).
• Detail Aggregation with the job-site assignment or doctor-patient treatment example.
• Present the comprehensive 7-point ER vs EER comparison table.