11 Chapter 11 Object-Oriented Databases Database Systems: Design, Implementation, and Management 4th Edition Peter Rob & Carlos Coronel

11 Object Schemas: The Graphical Representation of Objects Figure Shared Representation For All Objects Of the Class Person

11 State Of A Person Object Instance Figure 11.15

11 Defining Three Abstract Data Types Figure 11.16

11 Object Representation For Instances Of The Class Person With ADTs Figure 11.17

11 4The object space or object schema is the equivalent of a database schema or a relational schema at a given time. Object Schemas: The Graphical Representation of Objects Figure 11.18

11 Figure Referential Sharing Of Objects

11 4Class-Subclass Relationships Object Schemas: The Graphical Representation of Objects Figure Class Hierarchy

11 Employee Object Representation Figure 11.21

11 4Interobject Relationships: Attribute-Class Link u A attribute-class or interobject relationship is created when an object’s attribute references another object of the same or different class. u Relationship Representation: l Representing 1:M Relationships l Representing M:N Relationships l Representing M:N Relationships with an Intersection Class Object Schemas: The Graphical Representation of Objects

11 Class Hierarchy For The EDLP Retail Corporation Figure 11.22

11 4Representing 1:M Relationships u Related classes are enclosed in boxes to make relationships more noticeable. u The double line on the box’s right side indicates that the relationship is mandatory. u Connectivity is indicated by labeling each box. Object Schemas: The Graphical Representation of Objects

11 Representing The 1:M Relationship Figure 11.23

11 Representing The 1:1 And 1:M Relationships Figure 11.24

11 Employee-Dependent Relationship Figure 11.25

11 4Representing M:N Relationships Object Schemas: The Graphical Representation of Objects Figure 11.26

11 4Representing M:N Relationships with an Intersection Class Object Schemas: The Graphical Representation of Objects Figure Representing the M:N Relationship With Associated Attributes

11 Representing The M:N Relationship With Intersection Class Figure 11.28

11 Figure Object Space Representation

11 4Late and Early Binding: Use and Importance u With late binding the data type of an attribute is not known until execution time or run-time. u Late binding allows two different object instances of the same class to contain values of different data types for the same attribute. u Early binding allows the database to check the data type for each of the attribute’s values at compilation or definition time. Object Schemas: The Graphical Representation of Objects

11 Inventory Table With Predetermined (Base) Data Types Figure 11.30

11 Inventory Class With Early Binding Figure 11.31

11 OODM Inventory Class With Late Binding Figure 11.32

11 4Support for Versioning u Versioning is an OODM feature that allows us to track the history of change in the state of an object. u It is a very powerful modeling feature, especially in a CAD environment. Object Schemas: The Graphical Representation of Objects

11 OODM and Previous Data Models 4Object, Entity, and Tuple u An OODM object has additional characteristics such as behavior, inheritance, and encapsulation. u Such characteristics make OO modeling much more natural than E-R and relational modeling.

11 An Invoice Representation Figure 11.33

11 OODM and Previous Data Models 4Class, Entity Set, and Table u Class is a more powerful concept that allows not only the description of the data structure but also the description of the behavior. u A class allows both the concept and the implementation of abstract data types. 4Encapsulation and Inheritance u An object belonging to a class inherits all the properties of its superclasses. u Encapsulation hides the data representation and the method’s implementation from other objects and the user.

11 OODM and Previous Data Models 4Object ID u Object ID is not supported in either the E-R model or the relational model. u The hierarchical and the CODASYL models support some form of ID. 4Relationships u Relationships in an OODM can be of two types: interclass references or class hierarchy inheritance. u E-R and relational models use a value-based relationship approach.

11 OODM and Previous Data Models 4Access u E-R and relational models use an ad hoc, set- oriented query language. u OODM is suited to support both navigational and set-oriented access.

11 Object-Oriented DBMS Figure 11.34

11 Table 11.4 The Thirteen OODBMS Rules

11 How OO Affects Database Design 4OO database design approach provides both the data identification and the procedures or data manipulation to be performed. 4OO database design forces us to think of data and procedures as a self-contained entity. 4OO design is iterative and incremental in nature. 4DBA’s role is likely to change with more programming responsibilities. 4Lack of standards affects OO database design.

11 OODBMS: Advantages and Disadvantages 4Advantages u More semantic information. u Support for complex objects. u Extensibility of data types. u Improved performance with efficient caching. u Versioning. u Faster development and easy maintenance through inheritance and reusability. u Technology-driven product for next generation DBMS. u Potential to integrate DBMSs into a single environment.

11 4Disadvantages u Strong opposition from the established players. u Lack of theoretical foundation. u Retrogressive to the old pointer systems. u Lack of standard ad hoc query language. u Lack of business data design and management tools. u Steep learning curve. u Lack of resources. OODBMS: Advantages and Disadvantages

11 How OO Concepts Have Influenced the Relational Model 4New Features for Extended Relational (Object/Relational) Model u Extensibility of new user-defined (abstract) data types u Complex objects u Inheritance u Procedure calls (rules or triggers) u System-generated identifiers (OID surrogates)

11 How OO Concepts Have Influenced the Relational Model 4Philosophy that guides the relational model’s enhancements: u Semantic and object-oriented concepts are necessary to support the new generation of applications. u These concepts can and must be added to the relational model. u The benefits of the relational model must be preserved to protect the investment in relational technology and to provide downward compatibility.

11 The Next Generation of DBMS 4The next generation of DBMS is likely to incorporate features borrowed from: u Object-oriented database systems u Artificial intelligence systems u Expert systems u Distributed database u The Internet