Advanced Database Management

Faculty Name : Dr. Dipti Jadhav

Lecture 1

Chapter 3: Object Relational

and Extended Relational
Databases
 HISTORY OF DATABASES
file systems (1950s) store data after process created it
has ceased to exists
hierarchical/ network (1960s) •concurrency
• recovery
• fast access
• complex structures
relational (1970-80s) •more reliability
• less redundancy
• more flexibility
• multiple views
•ODBMS (1990s) better simulation
more (and complex) data types more
relationships (e.g. aggregation,
specialisation)
single language for database AND
programming
better versioning
no 'reconstruction' of objects
other OO advantages (reuse,
inheritance etc.)

Introduction to Testing
STONEBRAKER’S APPLICATION MATRI

Most applications will move to the upper right

Introduction to Testing
MOTIVATION

 Relational model (70’s):

 Clean and simple.
 Great for administrative and transactional data.
 Not as good for other kinds of complex data (e.g., multimedia, networks, CAD).

 Object-Oriented models (80’s):

 Complicated, but some influential ideas from Object Oriented
 Complex data types.

 Idea: Build DBMS based on OO model.

Programming languages have evolved from Procedural to Object Oriented. So why

not DBMSs ???

Introduction to Testing
RELATIONAL MODEL

 Relations are the key concept, everything else is around relations

 Primitive data types, e.g., strings, integer, date, etc.
 Great normalization, query optimization, and theory .

 What is missing??

 Handling of complex objects

 Handling of complex data types
 Code is not coupled with data
 No inherence, encapsulation, etc.

Introduction to Testing
RELATIONAL MODEL OF A ‘CAT’

Introduction to Testing
OBJECT ORIENTED MODEL OF A ‘CAT’

• Object-oriented database of a cat:

Introduction to Testing
 The first areas where ODBMS were widely used were:
 CASE
 CAD
 CAM

 Increasingly now used in:

 telecommunications
 healthcare
 finance
 multimedia
 text/document/quality management

Introduction to Testing
TWO APPROACHES

 Object-Oriented Model (OODBMS)

• Pure OO concepts

 Object-Relational Model (ORDBMS)

• Extended relational model with OO concepts

Introduction to Testing
DATABASE DESIGN PROCESS

Introduction to Testing
LOGICAL & PHYSICAL LAYERS

Introduction to Testing
EXAMPLE OF UML CLASSES

Introduction to Testing
FIRST APPROACH: OBJECT-ORIENTED MODEL

 Relations are not the central concept, classes and objects are the main concept.
 Object-Oriented DBMS(OODBMS) are DBMS based on an Object Oriented Data Model
inspired by OO programming languages.

 Main Features:
 Powerful type system
 Classes
 Object Identity
 Inheritance

 OODBMS are capable of storing complex objects, I.e., objects that are composed of other
objects, and/or multi-valued attributes.

Introduction to Testing
FEATURE 1: POWERFUL TYPE SYSTEM

 Primitive types
 Integer, string, date, Boolean, float, etc.

 Structure type
 Attribute can be a record with a schema

 Collection type
 Attribute can be a Set, Bag, List, Array of other types

 Reference type
 Attribute can be a Pointer to another object

Introduction to Testing
FEATURE 2: CLASSES

 A ‘class’ is in replacement of ‘relation’

 Same concept as in OO programming languages

 All objects belonging to a same class share the same properties and behavior

 An ‘object’ can be thought of as ‘tuple’ (but richer content)

 Classes encapsulate data + methods + relationships

 Unlike relations that contain data only

 In OODBMSs objects are persistency (unlike OO programming languages)

Introduction to Testing
FEATURE 3: OBJECT IDENTITY

 OID is a unique identity of each object regardless of its content.

• Even if all attributes are the same, still objects have different OIDs.

 Easier for references

 An object is made of two things:

 State: attributes (name, address, birthDate of a person)
 Behavior: operations (age of a person is computed from birthDate and current date)

Introduction to Testing
FEATURE 4: INHERITANCE

 A class can be defined in terms of another one.

 Person is super-class and Student is sub-class.

 Student class inherits attributes and operations of Person.

Introduction to Testing
STANDARDS FOR OBJECT-ORIENTED MODEL

 ODMG: Object Data Management Group (1991)

 provide a standard where previously there was none
 support portability between products
 standardize model, querying and programming issues

 Language of specifying the structure of object database

 ODL: Object Definition Language
 OQL: Object Query Language

 ODL is somehow similar to DDL (Data Definition Language) in SQL

Introduction to Testing
Overview of ODL & OQL
ODL: CLASSES & ATTRIBUTES

Introduction to Testing
ODL: RELATIONSHIPS

Introduction to Testing
ODL: RELATIONSHIPS & INVERSE RELATIONSHIPS

Introduction to Testing
ODL: MULTIPLICITY OF RELATIONSHIPS

Introduction to Testing
ODL: METHODS

Introduction to Testing
ODL: INHERITANCE

Introduction to Testing
ODL: INSTANCES & KEYS

 Instance of a class are all objects currently exist of that class

 In ODL that is called extent (and is given a name)
 Keys are not as important for referencing objects
 Because each object already has a unique OID
 Defining keys in ODL is optional
 ODL allows defining multiple keys (Comma separated)

Introduction to Testing
WHAT’S NEXT

 First Approach: Object-Oriented Model

• Concepts from OO programming languages

• ODL: Object Definition Language

• What about querying OO databases???

 OQL: Object Oriented Query Language

Introduction to Testing
OQL: OBJECT-ORIENTED QUERY LANGUAGE

 OQL is a query language designed to operate on databases described in ODL.

 Tries to bring some concepts from the relational model to the ODBMs
 E.g., the SELECT statement, joins, aggregation, etc.

 Reference of class properties (attributes, relationships, and methods) using:

 Dot notation (p.a), or
 Arrow notation (p->a)

 In OQL both notations are equivalent

Introduction to Testing
OQL: EXAMPLE QUERIES I

Introduction to Testing
OQL: EXAMPLE QUERIES II

Introduction to Testing
OQL OUTPUT

 Unlike SQL which produces relations, OQL produces collection (set, bag, list) of objects
 The object can be of any type

Introduction to Testing
OQL: AGGREGATION

Introduction to Testing
OQL: COLLECTION OPERATORS

Introduction to Testing
INTEGRATING OQL & EXTERNAL LANGUAGES

 OQL fits naturally in OO host languages

 Returned objects are assigned in variables in the host program

Introduction to Testing
WHAT’S NEXT

 First Approach: Object-Oriented Model:

 Concepts from OO programming languages

 ODL: Object Definition Language
 What about querying OO databases???
 OQL: Object Oriented Query Language

 Second Approach: Object-Relational Model

Introduction to Testing
SECOND APPROACH: OBJECTRELATIONAL MODEL

 Object-oriented model tries to bring the main concepts from relational model to the OO
domain

 The heart is OO concepts with some extensions.

 Object-relational model tries to bring the main concepts from the OO domain to the
relational model

 The heart is the relational model with some extensions

 Extensions through user-defined types

Introduction to Testing
CONCEPTUAL VIEW OF OBJECTRELATIONAL MODEL

 Relation is still the fundamental structure .

 Relational model extended with the following features

 Type system with primitive and structure types (UDT)
• Including set, bag, array, list collection types
• Including structures like records

 Methods
 Special operations can be defined over the user-defined types (UDT)
 Specialized operators for complex types, e.g., images, multimedia, etc

 Identifiers for tuples

 Unique identifiers even for identical tuples.

 References
 Several ways for references and de-references

Introduction to Testing
CONCEPTUAL VIEW OF OBJECTRELATIONAL MODEL

Allow of nested relations

 Repeating movies inside the stars

records is redundancy

 To avoid redundancy, use

pointers (references)

Introduction to Testing
SUPPORT FROM VENDORS

 Several major software companies including IBM, Informix, Microsoft, Oracle, and Sybase
have all released object-relational versions of their products

 Extended SQL standards called SQL-99 or SQL3

Introduction to Testing
SQL-99: QUERY LANGUAGE FOR OBJECTRELATIONAL MODEL

 User-defied types (UDT) replace the concept of classes

 Create relations on top of the UDTs

 Multiple relations can be created on top of the same UDT

Create Type <name > as (attributes and method declarations)

Introduction to Testing
CREATING UDT

Introduction to Testing
COLLECTIONS AND LARGE OBJECTS

 Book Type contains collections

 Arrays of authors -> capture the order of authors
 Set of keywords

Large object types:

CLOB: Character large objects book-review CLOB(10KB)

BLOB: binary large objects”

image BLOB(10MB)
movie BLOB(2GB)

Introduction to Testing
CREATING RELATIONS

 Once types are created, we can create relations

 In general, we can create tables without types
 But types provide encapsulation, inheritance, etc.

Introduction to Testing
CREATING RELATIONS

• A single primary key can be defined using Primary Key keyword .

• To reference another relation R, R has to be referenceable using REF keyword

Introduction to Testing
DEFINING RELATIONSHIPS

 One-to-many Or one-to-one
• Plug it inside the existing types

 Many-to-many
• Create a new type or new table referencing existing types

Introduction to Testing
WHAT’S NEXT

 First Approach: Object-Oriented Model

 Concepts from OO programming languages
 ODL: Object Definition Language
 What about querying OO databases???
 OQL: Object Oriented Query Language

 Second Approach: Object-Relational Model

 Conceptual view
 Data Definition Language (Creating types, tables, and relationships)
 Querying object-relational database (SQL-99)

Introduction to Testing
QUERYING OBJECT-RELATIONAL DATABASE

 Most relational operators work on the object-relational tables

 E.g., selection, projection, aggregation, set operations

 Some new operators and new syntax for some existing operators

 SQL-99 (SQL3): Extended SQL to operate on objectrelational database

Introduction to Testing
EXAMPLES I

Introduction to Testing
EXAMPLES II: DE-REFERENCING

Introduction to Testing
WHEN TO CONSIDER OODBMS OR ORDBMS

 Complex Relationships
 A lot of many-to-many relationships, tree structures or network (graph) structures.

 Complex Data
 Multi-dimensional arrays, nested structures, or binary data, images, multimedia, etc.

 Distributed Databases
 Need for free objects without the rigid table structure.

 Repetitive use of Large Working Sets of Objects

 To make use of inheritance and reusability

 Expensive Mapping Layer

 Expensive decomposition of objects (normalization) and recomposition at query time

Introduction to Testing
KEY BENEFITS OF ODBMS

 Persistence & Versioning

 Created objects are maintained across different database runs (persistent)
 Different evolving copies of the same object can be created over time (versioning)

Introduction to Testing
KEY BENEFITS OF ODBMS (CONT’D)

• Sharing in highly distributed environment

• Easier to share and distribute objects than tables

Introduction to Testing
KEY BENEFITS OF ODBMS (CONT’D)

 Better memory usage and less paging

 Bringing only objects of interest

Introduction to Testing
OBJECT-ORIENTED VS. OBJECTRELATION

 Object-oriented DBMSs
 Did not achieve much success (until now) in the market place
 No query support (Indexing, optimization)
 No security layer

 Object-relational DBMSs

 Better support from big vendors

 Tries to make use of all advances in RDBMSs
 Indexes, views, triggers, query optimizations, security layer, etc.
 • Work in progress --- Long way to go

Introduction to Testing
COMPARISON

Introduction to Testing