Semantic Data Models

1
Semantic Data Models

• David Nelson
• October 2001

2
Contents

• Introduction to Semantic Data Models
• Examples of Semantic Data Models
• The Functional Data Model
• Introduction to DAPLEX
• Data Definition
• Data Manipulation
• Derived Data
• System Considerations
• Further Reading

3
Introduction

• Relational Model has certain failings due to
  limited modelling capabilities
• Many models appeared which attempted to
  represent the real world more closely
• These are known as semantic data models
• Can be seen as the basis of object data models
  which we discuss in the following sessions

4
Semantic Data Models

• First semantic model was the E-R model by Chen
  in 1976
• However, this is not a true data model
• Why not?
• Data model structure rules manipulation
• Now a widely accepted modelling tool for
  relational database design
• But not a data model

5
Semantic Data Models

• In 1979, Codd devised RM/T
• an extended version of the relational model
• varying types of relations (one-to-many and
  subtyping)
• Other semantic data models
• Kuper and Vardi, Logical Data Model
• Hammer and McLeod, Semantic Data Model
• Su, Semantic Association Model
• Shipman, Functional Data Model
• We will look at Shipmans model

6
The Functional Data Model

• Devised by Shipman in 1981
• Defined functional data model with manipulation
  language DAPLEX
• No original implementation, but used in research
  projects
• ADAPLEX language based on ADA
• EFDM by Atkinson et. al. in Glasgow
• P/FDM by Gray, et. al. in Aberdeen
• Iris Object-Oriented Database

7
Introduction to DAPLEX

• DAPLEX is a database language which
  incorporates1
• A formulation of data in terms of entities
• A functional representation for both actual and
  virtual (derived) data relationships
• A rich collection of language constructs for
  expressing entity selection criteria
• A notion of subtype/supertype relationships among
  entity types

8
Quote!

• The DAPLEX language is an attempt to provide a
  database system interface which allows the user
  to more directly model the way he thinks about
  the problems he is trying to solve.

9
DAPLEX

• A data definition and manipulation language for
  database systems based on the functional data
  model
• Conceptually natural database interface language
• Basic constructs
• entity and function
• Derived functions
• Separate user views of the database

10
University Database

• declare student -gtgt entity
• declare name(student) -gt string
• declare dept(student) -gt department
• declare course(student) -gt course
• key_of student is name
• declare course -gtgt entity
• declare name(course) -gt string
• declare modules(course) -gtgt modules
• key_of course is name

11
Data Definition DECLARE statement

• Single Valued Functions
• declare name(student) -gt string
• declare dept(student) -gt department
• These functions return a singleton set
• Multi Valued Functions
• declare student -gtgt entity
• declare course(student) -gtgt course
• Multi Argument Functions
• declare grade(student, course) -gt integer

12
DECLARE Statement

• Function Inversion
• define course(instructor) -gt inverse of
  instructor(course)
• Inverse functions are automatically defined in
  P/FDM
• e.g. course_inv(instructor)
• Subtypes and supertypes
• declare person -gt entity
• declare student -gt person

13
Data Manipulation

• What are the names of all students taking EE
  courses from assistant professors?
• for each s in student
• such that some c in course(s) has
• name(dept(c))EE and
• rank(instructor(c)) ASSISTANT PROFESSOR"
• print(name(s))

14
Data Manipulation

• Aggregate operators include
• average, total, count, maximum, minimum
• Updating
• normally performed in P/FDM using a load utility
• Original definition of DAPLEX provided an update
  facility which allowed updating of both single
  and multi-valued functions

15
Derived Data DEFINE statement

• Defining new properties of objects based on the
  values of other properties
• e.g. define instructor(s in student) gt
• instructor(course(s))
• Can be used as primitive functions
• Similar to the concept of view
• Special operations
• inverse of, transitive of, intersection of, union
  of, difference of, compound of
• E.g. define enrolment() -gtgt compound of student,
  course(student)

16
Object Data Models

• Object models are attempts to provide models
  which more closely model the real world
• Developed to address the increasing complexity of
  database applications
• DAPLEX could be seen as a query language for
  object databases
• Object-oriented database system Iris uses form of
  DAPLEX as query language

17
Using P/FDM

• P/FDM is the version of DAPLEX we use in the
  university, based on Prolog
• You will be given a tutorial sheet showing how to
  use it
• To load a database, e.g. university database
• daplex(unidb.sch)
• pfdm_load(unidb.loa)
• daplex(unidb.d)
• Then daplex. enters the Daplex interpreter.

18
Further Reading

• Shipman, The Functional Data Model and the Data
  Language Daplex. ACM TODS, volume 6, no. 1,
  1981.
• Gray, Kulkarni, Paton, Object-Oriented Databases,
  1992