Relational model

0
Relational model
• Introduction
• Definitions
• References between relations
• Incomplete information
• Integrity constraints
• Primary key
• Tuple constraint and domain constraint
• Referential integrity constraint

1
Introduction
Relational model

2
Intuition
Courses
c1 M4880 Information systems Martin

c2
M2170 Computer network Smith

F0410 Databases Brown

c3

3
Relational model
• Proposed by E. F. Codd in 1970 to support higher abstract levels
compared to the previous models
• data independence
• Made available in commercial DBMSs in 1981,
• today it is the main model exploited in commercial DBMSs
• Based on the mathematical concept of relation
• each relation is represented in an informal way by means of a table

4
Example
Courses
Code Name TeacherID
M2170 Information systems D101
M4880 Computer Networks D102
F0410 Databases D321

Teachers ID Name Department Phone#

D101 Green Computer Engeneering 123456
D102 White Telecommunications 636363
D321 Black Computer Engeneering 414243

5
Definitions
Relational model

6
 Definitions

attributes
Attribute • Name of a table column

• The set of values that can be

Courses Code Name TeacherID
Domain
assumed by an attribute
M2170 Information systems D101
N-tuple
M4880 Computer Networks D102 • Table row
(or tuple)
F0410 Databases D321

• Number of n-tuples in a
Cardinality
relation
domain: set of identifiers
of the courses held at the
• Number of attributes in a
Politecnico Degree
relation

7
 Definitions
• Schema: describes the structure of the data
• virtually invariant over time
• is represented by the header of each table (table name and column names)
• Instance: consists of the contents of each table, i.e., the actual values
of the data
• variable over time, even very quickly
• is represented by the rows of the tables

Courses
Code Name TeacherID SCHEMA
M2170 Information systems D101

M4880 Computer Networks D102 INSTANCE

F0410 Databases D321

8
Properties
• Tuples (rows) are not ordered
• Tuples are distinct among them (there are no duplicated rows)
• Attributes are not ordered
• it is not possible to identify an attribute by means of its position

9
References between relations
Relational model

10
References between relations
• The relational model is value-based
• references between data in different relations are represented by means of
values of the domains
• Advantages:
• independent of the physical structure
• only information that is relevant from the application point of view is stored
• data can be easily transferred across different systems
• differently from pointers, the link is not oriented

11
Value-Based Reference: Example

Courses Code Name TeacherID

M2170 Information systems D101
M4880 Computer Networks D102
F0410 Databases D321

Teachers ID Name Department Phone#

D101 Green Computer Engeneering 123456
D102 White Telecommunications 636363
D321 Black Computer Engeneering 414243

12
Pointer-based reference: Example
Courses
Code Name TeacherID
M2170 Information systems
M4880 Computer Networks
F0410 Databases

Teachers ID Name Department Phone#

D101 Green Computer Engeneering 123456
D102 White Telecommunications 636363
D321 Black Computer Engeneering 414243

13
Null values
Relational model

14
Incomplete information
• Some information could be not available for all the tuples in the
relation
• Example:

• Student (StudentID, Surname, BirthDate, Phone#, DegreeYear)

• the phone number could be (temporarily?) unknown
• for students who have not yet graduated, year of degree is not defined

15
Null values
• To represent lack of information we could use a special value
belonging to the domain (0, empty string, 999, …)
• an “unused” value is required (example: DegreeYear=0, Phone#=?)
• “unused” values could become meaningful (Phone#= 999999)
• it is necessary to deal separately with “special” values ​in different applications
• This representation is not adequate

16
Null value
• Special value named null value (NULL)
• it is not a value of the domain
• it denotes both the absence of a domain value and a value not defined
• it must be used with caution (example: StudentID=NULL?)
• Notation: attributes that can have a null value (NULL) are often
highlighted with superscript * in the relation schema

ID Name Department Phone#*

D101 Green Computer Engeneering 123456
D102 White Telecommunications NULL
D321 Black Computer Engeneering 414243

17
Primary keys
Relational model

18
Unique identification for tuples

Students
StudentID Name Surname BirthDate EnrollementYear
64655 Mike Red 4/8/1978 1998
81999 Paul White 4/8/1978 1999
75222 Marco Red 8/3/1979 1998

• There are no pairs of students with the same value for the StudentID
• the StudentID uniquely identifies students
• There are no pairs of students with the same values for personal data
• name, surname and birth date uniquely identify students

19
Key
• A key is an attribute set that uniquely identifies tuples in a relation
• it is a property of the relational schema
• Formal definition: a set K of attributes is a key in a relation r if
• the relation r does not contain a pair of distinct tuples with the same values
for K (uniqueness)
• K is minimal (there exists no subset K’ of K that is still unique)

20
Example
• The attribute
{StudentID}
is unique and minimal, thus it is a key

• The attribute set

{Name, Surname, BirthDate}
is unique and minimal (none of its subsets is unique), thus it is a key

21
Superkey
• A set K of attributes is a key in a relation r if
• The relation r does not contain a pair of distinct tuples with the same values
for K (uniqueness)
• K is minimal (there are not proper subsets of K still unique)
• If only the first property is satisfied, K is a superkey of r

22
Examples
• The attribute set
{StudentID,Name}
is unique, but not minimal (the StudentID is unique), thus the
attribute set is a superkey, but it is not a key

• The attribute set

{BirthDate, EnrollementYear}
is unique and minimal: is it a general property?

23
Primary key
• If a key can assume the NULL value, it cannot be a key (the
uniqueness property is lost)
• it is mandatory to avoid the NULL values in the keys
• Solution:
• A reference key, which does not allow null values, is defined and set as
primary key
• The other keys (candidate keys) can assume null values
• References between data in different relations are defined by means of the
primary key

24
Primary key
• Notation: The attributes that make up the primary key are often
highlighted by underlining in the relation schema

ID Name Department Phone#*

D101 Green Computer Engeneering 123456
D102 White Telecommunications NULL
D321 Black Computer Engeneering 414243

25
Integrity constraints
Relational model

26
Integrity constraints
• Integrity constraint: property that must be satisfied by all correct
instances of the database
• Types of constraints:
• Intra-relational constraints
• defined on the attributes of a single relation (e.g.: uniqueness constraints, domain
constraints, and tuple constraints)
• Inter-relational constraints
• defined on multiple relation at the same time (e.g.: referential integrity constraints)

27
Uniqueness Constraints: Example

Courses Code Name TeacherID

M2170 Information systems D101
F0410 Computer Networks D102
F0410 Databases D321

Teachers ID Name Department Phone#

D101 Green Computer Engineering 123456
D102 White Telecommunications 636363
D321 Black Computer Engineering 414243

28
Domain Constraints
• Domain Constraint
• expresses conditions on the value assumed by a single attribute of a tuple
• it can be a Boolean expression (and, or, not) of simple predicates
• example: Grade > 0 and Grade 30

Exam CourseID StudentID Grade

M2170 S1234 23
M4880 S4321 28
F0410 s4321 40

29
Tuple constraint
• Tuple constraint
• expresses conditions on the values of each tuple, indipendently of other
tuples
• it can correlate many attributes
• it can be a Boolean expression (and, or, not) of simple predicates (e.g., comparison
between attributes, between an attribute and a constant)
• examples :
• Price = Cost + TaxPerc*Cost
• CumLaude = True if Grade = 30

CourseID StudentID Grade CumLaude

Exam
M2170 S1234 23 False
M4880 S4321 30 True
F0410 s4321 26 True
30
 Referential Integrity Constraint: Example
• Information in different relations is related through common values of one or
more attributes
• the TeacherID attribute in the Courses relation (referencing relation) refers to the
StudentID attribute in the Teacher relation (referenced relation)
• the values assumed by the TeacherID attribute in the Courses relation can only be
values taken by the ID attribute, the primary key of the Teacher relation
• the TeacherID attribute in the Courses relation is the foreign key of Courses

Foreign key
Courses (referencing relation) Teacher (referenced relation)
Code Name TeacherID ID Name Department Phone#
M2170 Information systems D101 D101 Green Computer Engineering 123456
M4880 Computer Networks D102 D102 White Telecommunications 636363
F0410 Databases D321 D321 Black Computer Engineering 414243
31
Referential integrity constraint
• Given two relations:
• R (referenced/external relation)
• S, that refers R through a set X of attributes (referencing/internal
relation)
values on a set X of attributes in a relation S can be exclusively
values for the primary key of the relation R
• The set X of attributes in S represents its foreign key
• Referential integrity constraints are imposed in order to guarantee
that the values in the referencing relation refer to actual values in
the referenced relation (the relational model is value-based)

32
 Referential Integrity Constraint: Example

Teacher ID Name Department Phone#

(referenced relation)
D101 Green Computer Engineering 123456
D102 White Telecommunications 636363
D321 Black Computer Engineering 414243

Foreign key

Courses (referencing relation)

Code Name TeacherID
M2170 Information systems D101
M4880 Computer Networks D102
F0410 Databases D101

33
 Referential Integrity Constraint: Example
Flight
(referenced relation)
F-ID Date

AZ111 16/10/2016

AZ234 4/12/2018
Foreign key
AZ543 9/3/2020

Ticket F-ID Date SeatNo Passenger

(referencing relation)
AZ111 16/10/2016 23 Luisa Reed

AZ111 16/10/2016 56 John White

AZ234 4/12/2018 9 Mark Black

AZ234 4/12/2018 11 Martha Black

AZ234 4/12/2018 21 Paul Austin

34
 Referential Integrity Constraint: Example
Flight
(referenced relation)
F-ID Date

AZ111 16/10/2016

AZ234 4/12/2018
Foreign key
AZ543 9/3/2020

Ticket F-ID Date SeatNo Persona

(referencing relation)
AZ111 16/10/2016 23 Luisa Reed

AZ111 16/11/2016 56 John White

AZ234 4/12/2018 9 Mark Black

AZ234 4/12/2018 11 Martha Black

AZ234 4/12/2018 21 Paul Austin

35
 Referential Integrity Constraint: Example
Flight
(referenced relation)
F-ID Date

AZ111 16/10/2016

AZ234 4/12/2018
Foreign key
AZ543 9/3/2020

Ticket F-ID Date SeatNo Passenger

(referencing relation)
AZ111 16/10/2016 23 Luisa Reed

AZ111 16/10/2016 56 John White

AZ234 4/12/2018 9 Mark Black

AZ234 4/12/2018 11 Martha Black

AZ543 4/12/2018 21 Paul Austin

36
Referential Integrity Constraint: Example

SP
Foreign keys
PId PName Color Size Store
P P1 Jumper Red 40 London SId PId Qty
P2 Jeans Green 48 Paris S1 P1 300
P3 Blouse Blue 48 Rome S1 P2 200
P4 Blouse Red 44 London S1 P3 400
P5 Skirt Blue 40 Paris S1 P4 200
P6 Shorts Red 42 London S1 P5 100
S S1 P6 100
S2 P1 300
SId SName #Employees City
S2 P2 400
S1 Smith 20 London
S3 P2 200
S2 Jones 10 Paris
S4 P3 200
S3 Blake 30 Paris
S4 P4 300
S4 Clark 20 London
S4 P5 400
S5 Adams 30 Athens

37
Referential Integrity Constraint: Example

SP
Foreign keys
PId PName Color Size Store
P P1 Jumper Red 40 London SId PId Qty
P2 Jeans Green 48 Paris S1 P1 300
P3 Blouse Blue 48 Rome S1 P2 200
P4 Blouse Red 44 London S1 P3 400
P5 Skirt Blue 40 Paris S1 P4 200
P6 Shorts Red 42 London S1 P5 100
S S1 P6 100
S2 P1 300
SId SName #Employees City
S2 P2 400
S1 Smith 20 London
S3 P2 200
S2 Jones 10 Paris
S4 P3 200
S3 Blake 30 Paris
S4 P4 300
S4 Clark 20 London
S9 P5 400
S5 Adams 30 Athens

38
Referential Integrity Constraint: Example

SP
Foreign keys
PId PName Color Size Store
P P1 Jumper Red 40 London SId PId Qty
P2 Jeans Green 48 Paris S1 P1 300
P3 Blouse Blue 48 Rome S1 P2 200
P4 Blouse Red 44 London S1 P3 400
P5 Skirt Blue 40 Paris S1 P4 200
P6 Shorts Red 42 London S1 P5 100
S S1 P7 100
S2 P1 300
SId SName #Employees City
S2 P2 400
S1 Smith 20 London
S3 P2 200
S2 Jones 10 Paris
S4 P3 200
S3 Blake 30 Paris
S4 P4 300
S4 Clark 20 London
S4 P5 400
S5 Adams 30 Athens

39