Chapter 3: RELATIONAL DATA MODEL

Lecturer: PhD. DO Thi Ngoc Diep

SCHOOL OF ELECTRICAL AND ELECTRONIC ENGINEERING
HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY

1
Review

• Entity relationship model?

• Entity, weak entity?
• Attributes, types of attributes?
• Relationships, types of relationships?

2
3.1. General concepts

• Relational data model uses set theory and first-order logic to

represent data.
• Suppose an object need to be stored in the database =>
• Define object attributes
• Define type and value domain for attributes

3
Relational model

• The current values of a relation are specified by a table

• Attribute values are (normally) required to be atomic (indivisible)
• The set of allowed values for each attribute is called the domain of the
attribute
• The special attribute value NULL = “unknown”
• An instance of relation <=> a tuple of attributes <=> a row/record in the
table

4
Relational model

• The current values of a relation are specified by a table :

• Cardinality of relation: number of rows/records of the table
• Degree of relation: number of columns/fields of the table
• None of two rows have the same set of values
• The order of columns/rows is not important

Degree = ?
Cardinality = ?

5
Relational model

• Comparison with ER model:

• Each table corresponds to a set of entities.
• Each column corresponds to an attribute.
• Each row is a set of real values of the a ributes, corresponding to an entity

Student

ID
Class
Full name Sex
Date of birth Address
6
3.2. Relation Schema

• A1, A2, …, An are attributes

• R = (A1, A2, …, An ) is a relation schema
• A relation r defined over schema R is denoted by r (R)

• Example:
• Full relation schema :
BOOK (id: integer, title: string, author: string, pub-year: integer)
• Short relation schema :
BOOK (id, title, author, pub-year)

7
For example

• BOOK(id, title, author, pub-year)

BOOK
Id Title Author Pub-year

1 The call of the wild Jack London 1903

2 The universe in a nutshell Stephen Hawking 2001

3 Hồng lâu mộng Tào Tuyết Cần 1791

• Mathematically: The relation schema R is the Cartesian product of the attributes value
domains :
Book = dom ( int ) × dom (string) × dom (string) × dom ( int )
dom() denotes the value domain

8
3.3. Constraints in relational model

• Constraint = rule to check data to satisfy certain conditions:

• Semantic constraints:
• Data belongs to a value domain, be smaller than 1 value, …
• Data belongs to a fixed format
• Key constraints
• Entity constraints
• Reference constraints

9
Key constraints

• None of two rows have the same set of values

• Superkey is used to uniquely identify a tuple in a relation.
• Given a relation R defined on the attribute set U = {A1, A2, …, An}.
• A subset K  U is called a super key of relation R if for any two tuples t1,
t2 in relation R: t1 [K]  t2 [K]

Superkeys: {Name, Salary, Eid} , {Eid}, {Name, Birthdate}

EMPLOYEE
Department Eid Name Birthday Salary CellPhone
D1 20001234 Nguyen Van A November 22, 30000 0912304101
1970
D1 20012322 Do Van Khoi February 1, 1978 27000 NULL
D2 20000980 Vu Minh April 30, 1967 50000 0903132123
D3 19991323 Nguyen Van B October 10, 50000 NULL
1966

10
Key constraints

Khóa tối thiểu

• Candidate key: K is a candidate key, if no subset of K can be a super

key: ! K' K where K' is a superkey
• A minimal super key, Removing any attribute causes the remaining
attributes to no longer be a super key
• Primary Key: The best candidate key.
• Each relation can have at most 1 primary key.
Superkeys : {Name, Salary, Eid}, {Eid}, {Name, Birthdate}
Candidate Key : {Eid}, {Name, Birthdate}
EMPLOYEE Primary key : {Eid}
Department Eid Name Birthday Salary CellPhone
D1 20001234 Nguyen Van A November 22, 30000 0912304101
1970
D1 20012322 Do Van Khoi February 1, 1978 27000 NULL
D2 20000980 Vu Minh April 30, 1967 50000 0903132123
D3 19991323 Nguyen Van B October 10, 50000 NULL
1966

11
Key constraints

• Foreign key: Phải tham chiếu đến 1 khóa chính của bảng khác

• One or several attributes which point/refer to tuples in another relation

• K is a foreign key of R1 that refers to relation R2 if K is a primary key of R2
• A set of attributes that belong to a relation, are not keys of that relation,
but are primary keys of another relation

Reference constraints

12
Entity constraints, Reference constraints

• Entity constraint: No attribute in the primary key can be assigned a

NULL value.
• Reference constraints by foreign keys:
• Identified between two relationships
• Maintains the association between tuples belonging to two relations
• If the foreign key is not null then there must be a tuple in the referenced
relationship where the primary key has the same value as the foreign key

13
Check Constraints

• Data in a relational database changes continuously over time.

• When there is an operation on the database => need to check the
constraints
• Add operations: data type constraints, semantic constraints, key constraints,
reference constraints,…
• Edit operations: data type constraints, semantic constraints, key constraints,
reference constraints,…
• Delete operations: reference constraints,…

14
 Check Constraints
ACCOUNT
CUSTOMER ACCOUNT HOLDER branchName balance accountNumber

customer Name Address customer account HaThanh 20000 C-12894349

Number Number Number DongDo 20000 C-12894350
111111 Anh Hai Ba Trung 111111 C-12894349 DongDo 3500 S-141510751
121314 Van Anh Hai Ba Trung 121314 C-12894350 HaThanh 50000 S-520522620
515016 Son Hoan Kiem 121314 S-141510751
BRANCH
515016 S-520522620
branchName address assets
111111 C-12894350
HaThanh Hai Ba Trung 900000000
DongDo Dong Da 400000000
ThangLong Hoan Kiem 500000000

1. Add Account (HaThanh, 5000, S-20071280) => check S-20071280 corresponds to which
customerNumber?
2. Add Account-Holder(12334, C-12894350)  check if there is customer 12334 ?
 Is there account C-12894350 ?
3. Edit Account ( C-12894350 , 20000, DongDo )  update BRANCH for DD, TL
to Account (C-12894350, 5000, ThangLong )
4. Delete Customer with customerNumber = '111111' => Delete/update in related tables
15
3.4. Transform ER to Relational Model

• Mapping of Regular Entity Types

• Mapping of Weak Entity Types
• Mapping of Binary 1:1 Relation Types
• Mapping of Binary 1:N Relationship Types
• Mapping of Binary M:N Relationship Types
• Mapping of Multivalued attributes
• Mapping of Ternary Relationship Types

16
3.4. Transform ER to Relational Model

• 1 entity set <=> 1 relation

• a ribute → column
• en ty → row/record/tuple
• Primary key of entity set → Primary key of relation

=> EMPLOYEE( EID , Name, Address, Birthdate, Salary)

17
3.4. Transform ER to Relational Model

• Transform composite attributes:

EMPLOYEE( Eid , Name, Address, …)

EMPLOYEE( Eid , Name, Number, Street, City, …)

18
3.4. Transform ER to Relational Model

• Transform multi-valued attributes

• Add a new entity

EMPLOYEE( Eid , Name, Address, Birthdate, Salary)

EMP_DEGREES( Eid , Degree )

19
3.4. Transform ER to Relational Model

• Transform weak entity

EMPLOYEE( EID , EName, Address, …)

EMP-DEPENDENT( EID , DName , Relationship, …)

20
3.4. Transform ER to Relational Model
From

1 1
• One-to-one Relationship Book set Design
Set: Dành cho
Book(bid, name, author, publisher, pub_date)
Design(dsid, color, shape)
• Both sides are Total participation : From
Book (bid , name, author, publisher, pub_date,
color, shape, from) // every book, every design 1 1
Book set Design
• One side is Total participation :
Book ( bid , name, author, publisher, pub_date , dsid, from) From
// every book is set a design
Design( dsid , color, shape ) // To keep also designs not for book 1 1
Book set Design

• Both side are Partial participations : From

Book( bid, name, author, publisher, pub_date )
Design( dsid , color, shape, bid, from ) // bid can be NULL !!! 1 1
Book set Design

Book( bid, name, author, publisher, pub_date )

Design( dsid , color, shape)
Book_Design( bid, dsid, from ) // or dsid as Primarykey, bid as foreign key. Need one more relation !!!

21
 3.4. Transform ER to Relational Model

• One-to-many relationship
1 N
Author write Book
Book( bid , name, publisher, pub_date )
Author( aid , name, address)

• Use foreign key

Book( bid , name, publisher, pub_date , aid )
Thêm 1 cột
2 cách Author( aid , name, address)
• Add a new relation
Book( bid , name, publisher, pub_date) Thêm 1 bảng mới
Author( aid , name, address)
Book_Author(bid, aid) // Need one more relation !!!

22
3.4. Transform ER to Relational Model

• Many-to-many relationship
• Add a new relation

n n
Student register Class

Student(s id , name, birthday)

Class( cid , name, room, teacher)
Registration( sid , cid )

23
3.4. Transform ER to Relational Model

• “ is-a” relationship
• Convert “parent” entity into a relation
• Convert “child” entity into a relation, containing only its own attributes
• Copy the key of the parent entity to the child entity

REGULAR_CALL
is-a
LONG_DISTANCE is-a CALL
is-a
CELL_CALL

24
3.4. Transform ER to Relational Model

REGULAR_CALL
is-a
LONG_DISTANCE is-a CALL
is-a
CELL_CALL

CALL ( Call_Identifier , Customer_Id , Source_Number , Destination_Number ,

TimeOfDay )

REGULAR_CALL ( Call_Identifier , Duration)

LONG_DISTANCE ( Call_Identifier , Duration, LongDistanceCarrier )
CELL_CALL ( Call_Identifier , Air_Time , Land_Time )

25
3.4. Transform ER to Relational Model

• Ternary Relationship
• Create a new relationship and add primary key
• 1:1:1 type
• 3 candidate keys:(Aid, Bid), (Aid, Cid), (Bid, Cid)

• 1:1:N type
• 2 candidate keys: (Aid, Bid), (Cid, Bid)

• 1:N:N type
• 1 candidate key: (Aid, Bid)

• N:N:N type
• 1 candidate key: (Aid, Bid, Cid)

26
Schema for a database

• A relational model is represented as a diagram: the logical

structure of the database

Database System Concepts – Siberschatz, Korth, Sudarshan

27
Exercise

• Teams

28