| Logical
Data Modeling Presented by
Duncan Dwelle
Prepared for
This Copy is not Final!
Before we finish -
Our Proof of Sound Methods:
Why Build a Model?
Why Build a Model?
Were Does a Model Start?
Were Does a Model Start?
Were Does a Model Start?
A Model Must Serve the Business
An Impedance Mismatch
Relational Layer
Relational Layer
RDBMS Layer
Logical Model
Multi-Model Layers
Different Terms @ Same Idea
Different Terms for the Relation
Different Terms for the Domain
Different Terms for the Join
Different Terms for the Key
One Theory - Two Models
Schema vs. Logical Model
Schema vs. Logical Model
Schema vs. Logical Model
Schema vs. Logical Model
Schema vs. Logical Model
Schema vs. Logical Model
Schema =|¹ Logical Model
Logical Model with Sub-Typing
Denormalization
Schema: No Sub-Type Tables
Logical Model of Large Objects
Vertical Table Split Schema
Horizontal Table Split Schema
Other Transformation Needs
Schema =|¹ Logical Model
One Logical Model into Many -
A Logical Model Provides:
The Logical Model’s Purpose:
How do We Use Models?
Conflicting Goals and Uses
What is in the Logical Model?
Structure is -
Constraints are -
Rules govern -
Sound Structure is Normalized
Constraints are Crucial to
Meaning
Domains are Constraints
Relationships are Constraints
Constraints Must Anticipate
Structure Implies Constraints
Rules Assign Values
When Structure Includes Rules
Foreign Key Implies Cardinality
Foreign Key Implies No Value
Forced Values
Child Implies No Cardinality
Limiting Unruly Children
Relationship ¹ Rule
Data Model ¹ Database Model!
The Logical Model’s Limits:
Your Dilemma - Option A:
Your Dilemma - Option B:
Our Ideal - Object Data Design
AIS - Your Expert Source
Contact Duncan Dwelle
|
