Relational and

UNIT 2 RELATIONAL DATABASE E-R

Structure Page Nos.

2.0 Introduction
2.1 Objectives
2.2 The Relational Model
2.2.1 Domain, Attribute, Tuple and Relation
2.2.2 Super keys Candidate keys and Primary keys for the Relations
2.3 Relational Constraints
2.3.1 Domain Constraint
2.3.2 Key Constraint
2.3.3 Integrity Constraint
2.3.4 Update Operations and Dealing with Constraint Violations
2.4 Relational Algebra
2.4.1 Basic Set Operation
2.4.2 Cartesian Product
2.4.3 Relational Operations
2.5 Summary
2.6 Solution/Answers

2.0 INTRODUCTION
In the first unit of this block, you have been provided with the details of the Database
Management System, its advantages, structure, etc. This unit is an attempt to provide
you information about relational model. The relational model is a widely used model
for DBMS implementation. Most of the commercial DBMS products available in the
industry are relational at core.
Relational model is based on theory of relations and was first proposed by E.F. Codd.
In this unit we will discuss the terminology, operators and operations used in
relational model.

2.1 OBJECTIVES
After going through this unit, you should be able to:

 describe relational model and its advantages;

 perform basic operations using relational algebra;

 identify the relational constraints in a database system;

 identify key of a relation.

2.2 THE RELATIONAL MODEL

A model of database system basically defines the structure or organisation of data,
the set of integrity constraints on the data and a set of operations that can be
performed on the data. The basic structure of data in a relational model in the form of
two-dimensional tables. A table in this model is called a relation. Such organization of
data simplifies enforcement of integrity constraints and database operations. The
following table represents a simple relation:

1
The Database
PERSON_ID NAME AGE ADDRESS
Management System
Concepts 1 Sanjay Prasad 35 B-4,Modi Nagar, UP
2 Sharada Gupta 30 Pocket 2, Mayur Vihar, Delhi
3 Vibhu Datt 36 C-2, Saket, New Delhi
Figure 1: A Sample Person Relation

Following are some of the advantages of relational model:

 Ease of use: The simple tabular representation of database helps the user define
and query the database conveniently. For example, you can easily find out the
age of the person whose first name is “Vibhu”.
 Flexibility: Since the database is a collection of tables, new data can be added
and deleted easily. Also, manipulation of data from various tables can be done
easily using various basic operations. For example, you can add a telephone
number field in the table at Figure 1.
 Accuracy: In relational databases the relational algebraic operations are used to
manipulate database. These are mathematical operations and ensure accuracy
(and less of ambiguity) as compared to other models. These operations are
discussed in more detail in Section 2.4.
2.2.1 Tuple, Attribute, Domain and Relation
Before we discuss the relational model in more detail, let us first define some very
basic terms used in this model.
Tuple: Each row in a table represents a record or information of a single object,
which is also termed as a tuple. For example, Figure 1 has three records or tuples. A
record/tuple consists of a number of attributes, which is defined next.
Attribute: A relation consists of large number of columns. Each of these columns,
which defines a separate data value, is termed as an attribute. The column name in a
relation is generally related to the meaning of data items of that column. For example,
Figure 2 represents a relation PERSON. The columns PERSON_ID, NAME, AGE,
ADDRESS and TELEPHONE are the attributes of the relation PERSON and each
row in the relation represents a separate tuple (record).
Relation Name: PERSON
PERSON_ID NAME AGE ADDRESS TELEPHONE
1 Sanjay Prasad 35 B-4,Modi Nagar, UP 011-25347527
2 Sharada Gupta 30 Pocket 2, Mayur 023-12245678
Vihar, Delhi
3 Vibhu Datt 36 C-2, Saket, New 033-1601138
Delhi
Figure 2: An extended PERSON relation

The relation of Figure 2 consists of 5 attributes, therefore, each tuple in this relation is
called a 5-tuple. Thus, if a relation has n attributes, then each record in that relation
would be termed as n-tuple.
Domain: A domain is a set of permissible values that can be accepted by a specific
attribute. For example, in Figure 2, if you assume that there may be a maximum of
100 persons in the relation, then you may assign PERSON_ID to a domain of integer
values, which should be in the range from 1 to 100 only. Once you have assigned this
domain, then you will not be able to assign any value below 1 and above 100 to
PERSON_ID. The domain of attribute AGE can be integer values between 0 and 150.
The domain can be defined by assigning a type or a format or a range to an attribute.
For example, a domain for a number 501 to 999 can be specified by having a 3-digit
number format having a range of values between 501 and 999. Domains need not be
contiguous numbers. For example, the enrolment number of IGNOU has the last digit
as the check digit, thus the enrolment numbers are non-continuous.

2
Relation: Each table is a relation. A relation is defined using two basic aspects, viz. Relational and E-R
schema and an instance. Models

Relational Schema: A relational schema denoted as R, specifies the relation’s name,

the list of attributes of the relation and the domain of each attribute, which is denoted
as R (A1, A2, ..., An). The relation R has n attributes, which is also called the degree of
a relation.
For example, the relation schema of the relation PERSON (see Figure 1) can be
defined as the follows:

PERSON (PERSON_ID: Integer, NAME: Character, AGE: Integer,

ADDRESS: Character)

Since the PERSON relation contains four attributes, so this relation is of degree 4.

Relation Instance or Relation State: A relation instance denoted as r is a collection of

tuples for a given relational schema at a specific point of time.

A relation state r of the relation schema R (A1, A2, ..., An), also denoted by r(R) is a
set of n-tuples

r = {t1,t2,…............tm} has m tuples

Where each n-tuple is an ordered list of n
values t = <v1,v2,… , vn>
where each vi belongs to domain (Ai) or contains null values.

Let us elaborate the definitions above with the help of examples:

Example 1:
RELATION SCHEMA For STUDENT:
STUDENT (RollNo: string, name: string, login: string, age: integer)
RELATION INSTANCE

STUDENT
ROLLNO NAME LOGIN AGE
t1
3467 Shikha [email protected] 21
t2 4677 Kanu [email protected] 20

Where t1 = (3467, shikha, [email protected], 20) for this relation instance, the
number of tuples m = 2 and each tuple contains n = 4 values.

Example 2:
RELATIONAL SCHEMA For PERSON:
PERSON (PERSON_ID: integer, NAME: string, AGE: integer, ADDRESS:
string)
RELATION INSTANCE
In this instance, the number of tuples m = 3 and each tuple has n = 4 values.

PERSON_ID NAME AGE ADDRESS

1 Sanjay Prasad 35 B-4,Modi Nagar, UP
2 Sharad Gupta 30 Pocket 2, Mayur Vihar, Delhi
3
3 Vibhu Datt 36 C-2, Saket, New Delhi

4
The Database
Management System
Concepts If a relation has proper integrity constraints, then each tuple of that relation would be
valid tuple. It may be noted that ‘Null’ value can be assigned for some of the attributes
where the values are unknown or missing. However, the Null’ value cannot be
assigned to certain attributes, this will be explained in the later sections.
Ordering of tuples
In a relation, tuples are not inserted in any specific order. Ordering of tuples is not
defined as a part of a relation definition. However, records may be organised later
according to some attribute value in the storage system. For example, records in
PERSON table may be organised according to PERSON_ID. Such data organisation
depends on the requirement of the underlying database application. However, for the
purpose of display, you may get these tuples displayed in the sorted order of age. The
following table is sorted by age. Please note that this is sorted related is same as that
or relation displayed in the order of PERSON_ID as each and every tuple is same. It is
also worth mentioning that relational model does not allow duplicate tuples.

PERSON

PERSON_ID NAME AGE ADDRESS

2 Sharad Gupta 30 Pocket 2, Mayur
Vihar, Delhi
1 Sanjay Prasad 35 B-4,Modi Nagar, UP
3 Vibhu Datt 36 C-2, Saket, New
Delhi

2.2.2 Super Keys, Candidate Keys and Primary Keys for the
Relations
As discussed in the previous section, ordering of relations does not matter and all
tuples in a relation are unique. However, can you uniquely identify a tuple in a
relation? To answer this question, let us discuss the concept of keys in relations.
Super Keys
A super key is an attribute or set of attributes used to identify the records uniquely in
a relation.
For Example, in the Relation PERSON described earlier PERSON_ID is a super key
since PERSON_ID is unique for each tuple/record. Similarly (PERSON_ID, AGE)
and (PERSON_ID, NAME) are also super keys of the relation PERSON since their
combination is also unique for each tuple/record.
Candidate keys
Super keys of a relation may contain extra attributes. A candidate key is a minimal
super key, which means that a candidate key does not contain any extraneous
attribute. An attribute is called extraneous if even after removing it from the key, the
remaining attributes still has the properties of a super key. A relation may have several
candidate keys. A candidate key may contain one or many attributes of a relation.
The following properties must be satisfied by a candidate key:
 In any instance of a relation, the value of the candidate key attribute(s) should be
unique.
 You cannot put a ‘Null’ value in any attribute that is a part of the candidate key.
This rule is also termed as the entity integrity rule. Thus, a candidate key is unique
and not null.
 A candidate key should have a minimal set of attributes.

5
 The value of a candidate key must be stable, which means it should not change Relational and E-R
frequently or its value change should not be outside the control of the system. Models

A relation can have more than one candidate keys and one of them can be chosen as a
primary key.

For example, in the relation PERSON the two possible candidate keys are
PERSON_ID and NAME (assuming unique names exists in the table). PERSON_ID
may be chosen as the primary key.

Check Your Progress 1

Answer the following questions for the relational instance s of the following
Supplier relation S:
S
SNO SNAME CITY
S1 Smita Delhi
S2 Jim Pune

1. What are the different attributes of relation S and how many tuples
does S have?
………………………………………………..
2. What are the domains of the attributes of S?
……………………………………………….
…………………………………………..…….
3. On sorting this relation on the field CITY, will the relation change?
Will the order of tuples in the relation change?
……………………………………………………………………..
4. List the super keys, all the possible candidate keys and the primary key
of the relation
…………………………………..

2.3 RELATIONAL CONSTRAINTS

A relational database is collection of relations. Each relation consists of tuples, which
consists of attributes. You can associate constraints, called the relational constraints,
with the attributes. Such constraints can be of the following types::
 Domain Constraints
 Key Constraint
 Integrity Constraints

2.3.1 Domain Constraint

These constraints bind the possible values of attribute in a relation to a specific set of
values, called the domain of the attribute. For example, an attribute COUNTRY may
have domain of names of all the possible names of the countries. In commercial
relational database management systems, the data type associated with an attribute
defines the broad domain of an attribute. These domains include:
1) Integral Data type like integer, long etc.
2) Data types related to real numbers like float, double etc.
3) Data types relating to alphabets like characters, strings etc.
4) Boolean

6
The Database
Management System
Thus, domain constraint specifies the possible set of values that you want to put in an
Concepts attribute of a relation. The values that appear in each attribute/column must be drawn
from the domain associated with that column.
For example, consider the relation:
STUDENT

ROLLNO NAME LOGIN AGE

3467 Shikha [email protected] 21
4677 Kanu [email protected] 20

In the relation above, AGE of the relation STUDENT always belongs to the integer
domain within a specified range (say 0 representing just born to 150), and not to
strings or any other domain. Within a domain non-atomic value should be avoided.
This sometimes cannot be checked by domain constraints. For example, a database
which has area code and phone numbers as two different fields will take phone
numbers as-

Area_code Phone
11 29534466

A non-atomic value in this case for a phone can be 1129534466, however, this value
can be accepted by the Phone field only.
2.3.2 Key Constraint
This constraint states that the key attribute value in each tuple must be unique, i.e., no
two tuples can contain the same value for the key attribute. This is because the value
of the primary key is used to identify a unique tuple in a relation.

Example 3: If A is the key attribute in the following relation R, then A1, A2 and A3
must be unique.

R
A B
A1 B1
A3 B2
A2 B2

Example 4: In relation PERSON, PERSON_ID is the primary key so PERSON_ID

cannot be assigned same for two different persons.
2.3.3 Integrity Constraint
There are two types of integrity constraints:
 Entity Integrity Constraint
 Referential Integrity Constraint

Entity Integrity Constraint:

It states that any attribute of a primary key cannot take a Null value. This is because
the primary key is used to identify individual tuple in the relation. You will not be
able to identify the records uniquely if they contain null values for the primary key
attributes. This constraint is specified for each relation of a database.
7
 Relational and E-R
Models
Example 5: Let R be a relation; an instance r of R is given below. Is the instance r
valid?

A# B C
Null B1 C1
A1 B2 C2
Null B2 C3
A2 B3 C3
Null B1 C5

Note: A ‘#’ in the headings row is indicating that A is the Primary key of R.

In the instance r of relation R above, the primary key has Null values in the tuple
t1, tuple t3 and tuple t5. As per the entity integrity constraint, Null value in primary
key is not permitted. Thus, relation instance r is an invalid instance.
Referential integrity constraint
For defining the referential integrity constraint, first we explain the concept of a
foreign key and foreign key constraint.
Consider an attribute set A of a relation R, which references an attribute set B in a
relation S, then A will be referred to foreign key in R provided it fulfills the following
conditions:
1. B is the Primary key of S.
2. A and B are defined over the same domains. A is called the foreign key
in relation R, which references B in relation S.
3. For every value x of attribute set A, in any instance r of R, there exist a
tuple in any instance s of S, where the value of attribute set B is x. Please
note that there will be only one such tuple having the value x, as B is the
Primary key of S. This is called foreign key constraint.
4. Please note that the relation R is a referencing relation and S is
called referenced relation.

A referential integrity constraint is more generic form of foreign key constraint, which
requires that the attribute of the referencing relation must be present in the attribute of
referenced relation in some tuple. Thus, foreign key constraint is a specific form of
referential integrity constraint. Foreign key constraint is explained with the help of
following example.
Example 6: Consider the two relational instance r of relation R and s of relation S.
Now answer the following questions:
a) If C in R is foreign key of C in S, then is the following instances r and
s valid?
b) A new tuple having values (A6, B2, C4) is added in r. Does it
violate foreign key constraint?

Instance r of R Instance
E s of S
A# B C^ C#
A1 B1 C1 E1 C1
A2 B2 C2 E2 C3
A3 B3 C3 E3 C5
A4 B4 C3 E2 C2
A5 B1 C5

8
The Database
Management System
Notes:
Concepts
 ‘#’ identifies the Primary key of a relation.
 ‘^’ identifies the foreign key in a relation.

a) In the example above, every value of C in r is matching with the values of C

in s in some tuple. The r contains the values C1, C2, C3, C5 and s contains all
these values. Thus, the instances r and s does not violate foreign key
constraint.
b) If you try to insert the tuple (A6, B2, C4) in r, then it is violating the foreign
key constraint for the present instances of relations, as there is no tuple in s,
which contains value C4 in s.

2.3.4 Data Updating Operations and Constraint Violations

There are three basic data updating operations to be performed on relations:
 Insertion of tuples
 Deletion of tuples
 Update the value of attribute(s) in a relation
The INSERT Operation:

 To add new records in a database system, you use insert operation. For example,
in the instance r of R of example 6, you may like to add a new record <’A5’, 25,’
C6’>. Addition of a new record may cause constraint violation, which are
explained below:

 Violation of Domain constraint: Such a violation occurs if the domain of a value,

which you are trying to insert in an attribute, does not match the attribute’s
domain. e inserting a numeric value 25 into attribute B, whose domain is
alphanumeric sting. This will cause a domain constraint violation.
 Violation of Key constraint: This violation occurs when you try to insert a key
value, which already exists in some tuple of existing relational instance. For
example, when you are trying to insert tuple <’A5’, 25, ’C6’> in R, you are
inserting a value A5 in the key attribute A. However, the value A5 already exists
in attribute A of tuple t5. Therefore, it is a violation of key constraint.
 Violation of Entity Integrity constraint: This violation will occur; in case you try
to insert a Null value into the primary key attribute. For example, if you try to
insert a tuple <’Null, 25, ’C6’> in R, you are violating this constraint.
 Violation of Referential Integrity constraint: Consider that while inserting a new
tuple, the value that you are trying to insert in a foreign key attribute, which refers
to an attribute in another relation where it is the primary key, does not match with
any value of referred attribute of the referenced relation. (Please see example 6 par
(b)). For example, when you are trying to insert tuple <’A5’, 25, ’C6’> in R, you are
inserting a value C6 in the foreign key attribute A of R. However, the value C6 is
not a value in referred attribute C in referenced relation S. Thus, insertion of this
tuple violates the referential integrity constraint.
Dealing with constraints violation during insertion:
If the insertion violates one or more constraints, then two options are available:

 Default option: - Insertion can be rejected and the reason of rejection can also
be explained to the user by DBMS.
 Ask the user to correct the data, resubmit, also give the reason for rejecting the
insertion.

9
Example 7: Relational and E-R
Models
Consider the Relation PERSON of Example 2:

PERSON

PERSON_ID NAME AGE ADDRESS

1 Sanjay Prasad 35 B-4,Modi Nagar, UP
2 Sharad Gupta 30 Pocket 2, Mayur Vihar, Delhi
3 Vibhu Datt 36 C-2, Saket, New Delhi

(1) Insert <1, ‘Vipin’, 20, ‘Mayur Vihar’> into PERSON

Violated constraint: - Key constraint
Reason: - Primary key 1 already exists in PERSON.
How to resolve: - DBMS could ask the user to provide valid PERSON_ID value
and accept the insertion if valid PERSON_ID value is provided.

(2) Insert <’null’, ‘Anurag’, 25, ‘Patparganj’> into PERSON

Violated constraint: - Entity Integrity constraint
Reason: - Primary key is ‘null’.
How to resolve: - DBMS could ask the user to provide valid PERSON_ID value
and accept the insertion if valid PERSON_ID value is provided.

(3) Insert <’abc’, ‘Suman’, 25, ‘IP college’> into PERSON

Violated constraint: - Domain constraint
Reason: - value of PERSON_ID is given a string which is not valid.
How to resolve: - DBMS could ask the user to provide valid PERSON_ID value
and accept the insertion if valid PERSON_ID value is provided.

(4) Insert <10, ‘Anu’, 25, ‘Patpatganj’> into PERSON

Violated constraint: - None

Note: In the first 3 cases of constraint violations above DBMS will reject the insertion.

The Deletion Operation:

The delete operation is used to delete some existing records from a relation. To delete
some specific records from the database a condition is specified based on which
records are selected for deletion.

Constraints that can be violated during deletion

Only one type of constraints can be violated during deletion, it is referential integrity
constraint. When you want to delete a record in a referenced relation, there is a
possibility that you may delete a tuple whose attribute is being referenced by the
referencing relation, where the attribute is the foreign key. Please go through the
example 8 very carefully.

Dealing with Constraints Violation

If the deletion violates referential integrity constraint, then three options are available:
 Default option: - Reject the deletion. It is the job of the DBMS to explain to
the user why the deletion was rejected.
 Attempt to cascade (or propagate) the deletion by deleting tuples whose
foreign key references the tuple that is being deleted.
 Change the value of referencing attribute that causes the violation.
1
The Database
Management System Example 8:
Concepts Let instance r of relation R be:

A# B C^
A1 B1 C1
A2 B3 C3
A3 B4 C3
A4 B1 C5

And instance q of relation Q be:

C# D
C1 D1
C3 D2
C5 D3

Note:
1) ‘#’ identifies the Primary key of a relation.
2) ‘^’ identifies the Foreign key of a relation.

(1) Delete a tuple with C# = ’C1’ in Q.

Violated constraint: - Referential Integrity constraint is violated
Reason: - If stated tuple is deleted from Q, the first tuple of R, which contains
the Foreign key (C) value C1will violate the foreign key constraint.
How to resolve: - Options available are
1) Reject the deletion.
2) DBMS may automatically delete all tuples from relation Q and S with C #
= ‘C1’. This is called cascade detection.
3) The third option would result in putting NULL value in R where the value
of attribute C is C1, which is the first tuple of R.

The Update Operations:

Update operations are used for modifying values of attributes in a database. The
constraint violations faced by this operation are logically the same as the problem
faced by Insertion Deletion Operations. You may define these actions yourself.

Check Your Progress 2

1 Consider the tables Suppliers, Parts, proJect and SPJ relation instances in the
relations below.

S P
SNO SNAME CITY PNO PNAME COLOUR CITY
S1 Smita Delhi P1 Nut Red Delhi
S2 Jim Pune P2 Bolt Blue Pune
S3 Ballav Pune P3 Pen White Mumbai

11
 J SPJ Relational and E-R
JNO JNAME CITY SNO PNO JNO QUANTITY Models

J1 Sorter Pune S1 P1 J1 200

J2 Display Mumbai S2 P2 J2 700

Using the instance of relations as given above, answer the following questions:

1 List a domain constraint for S.

………………………………………………………………………
………………………………………………………………………
2 List the Primary keys of all the relations and primary key constraint for
the SPJ relation.
………………………………………………………………………
………………………………………………………………………
3 List all the Foreign keys and Foreign key constraint.
………………………………………………………………………
………………………………………………………………………
4 What would be the constraint violations if the following operations
are performed:
(a) Insert <Null, ‘Jack’, ‘Mumbai’> into S
(b) Insert <’S2’, Null, ‘J3’, 200> into SPJ
(c) Insert <’P2’, ‘Pencil’, ‘Grey’, ‘Kolkata’> into P
(d) Insert <’J3’, ‘Monitor’,’ Jaipur’> into J
………………………………………………………………………
………………………………………………………………………
………………………………………………………………………

2.4 RELATIONAL ALGEBRA

Relational Algebra is a set of basic operations used to manipulate the data in

relational model. These operations enable the user to specify basic retrieval requests.
The result of retrieval is a new temporary relation, which is computed after
performing these operations on one or two relations. Some of these operations can be
classified in three categories:

 Relational Operations
These can be unary operations
1. SELECTION
2. PROJECTION
Or binary operations
1. CARTESIAN PRODUCT
2. JOIN
 Basic Set Operations
1. UNION
2. INTERSECTION
3. SET DIFFERENCE
 Assignment Operation, rename and other operations

In this section, we will discuss the relational operation and basic set operations. You
may refer to the other operations from the further readings.

2.4.1 Relational Operations

Relational algebra forms the basis of the query languages on a database system. Thus,
by studying relational algebra you may be able to write better queries for a database
system. Let us discuss these operations in detail.
1
The Database
Management System
Concepts Selection Operation:

Selection is a urinary operator, that is used to choose only those tuples from a relation
that fulfil a given criteria. It is represented using the mathematical symbol , as given
below:

 <Selection condition> (Relation)

You should specify a Boolean expression as a <Selection condition>. A selection

condition uses comparison operators like { , , ,=, <, <} and logical connectors like
and (), or () and not ().

Example 13:
Consider the relation PERSON. If you want to display details of persons having age
less than or equal to 30, then the selection operation will be used as follows:
 AGE <=30 (PERSON)

The resultant relation will be as follows:

PERSON_ID NAME AGE ADDRESS

2 Sharad Gupta 30 Pocket 2, Mayur Vihar, Delhi

Note:

1) Selection operation is commutative, i.e.,

 <condition1> (<condition2> (R)) =  <condition2> (<condition1> (R))
Hence, you can apply a sequence of Selection operations in any order

2) You can apply more than one conditions by using logical connectors.

Projection operation

The projection operation is used to select specific attributes from amongst all the
attributes of records. This is denoted as  .

 List of attributes for projection (Relation)

Example 14:

Consider the relation PERSON. If you want to display only the names of persons, then
the projection operation will be used as follows:

 NAME(PERSON)

The resultant relation will be as follows:

NAME
Sanjay Prasad
Sharad Gupta
Vibhu Datt

Please note that for projection operation:

 <List1> (  <list2> (R))=  <list1> (R) provided <list2> contains attributes in <list1>.

13
2.4.2 Cartesian Product Relational and E-R
Models
Given two relations R and S, the cartesian product of these two relations can be
represented as T = R S. The cartesian product operation produces all possible
combinations of the tuples of tuples of relations R and S. The cartesian product
operation is commutative as well as associative. In addition, the degree of the
resultant relation (T) is the sum of degree of relation R and relation S. The Cartesian
product can be expressed using the following mathematical expression:
T= {t1 || t2 | t1 R1  t2  R2}. Here, || is a concatenation operator.

Example 12: Consider the following two relational instances of relation R1 and R2.
What would be the cartesian produce of the relations?
C
R1 R2
A B
A1 B1 C1
A1 B2 C2
A2 B2
A2 B3

The Cartesian produce of the relations is denoted by R3 = R1 R2. The relation R3

will be as shown below:

A B C
A1 B1 C1
A1 B1 C2
A1 B2 C1
A1 B2 C2
A2 B2 C1
A2 B2 C2
A2 B3 C1
A2 B3 C2

Please note that in cartesian product every tuple/record of R1 is concatenated with

every record of R2. You can observe that the record <A1, B1> is concatenated with
both the records of R2, that is why in the output R3 you have tuples < A1, B1, C1>
and <A1, B1, C2>, likewise for all the tuples of R1. Please also note that R1 has 4
tuples and R2 has 2 tuples, therefore, R3 has 42=8 tuples.

The JOIN operation

JOIN is a binary operator. It combines two relations based on a given join condition.
A JOIN operation is represented by mathematical symbol ø. But how does JOIN
operation will combine two relations? It would require that there should be at least
one attribute in each of the relation, which have the same domain. Such attributes are
called domain compatible attribute.

Syntax:

R1ø<join condition>R2 is used to combine related tuples from two relations R1 and R2 into
a single tuple.
<join condition> is of the form:
<condition>AND<condition>AND….............AND<condition>.
 Degree of Relation:
Degree (R1ø<join condition>R2) <= Degree (R1) + Degree (R2).

1 35
The Database
Management System
Concepts  Three types of joins are there:

a) Theta () join

When each condition is of the form A  B, where A is an attribute of R1 and B is

an attribute of R2; both A and B have the same domain; and  is one of the
comparison operators { , , ,=, <, <}.

b) Equijoin
Equijoin is a restricted form of When each condition appears with equality
operator (=) only.

c) Natural join
Natural join is defined as a specialised type of join, when the joining attributes in
the two joining relations have the identical name, in addition to the identical
domain and the condition for the join operation is equality (=) condition. In such
cases only one joining attribute is kept in the result.
The following example explains the Natural join operation.
Example 15:
Consider the instance of the following relations. The primary key of STUDENT
relation is ROLLNO and foreign key is COURSE_ID, which references the primary
key COURSE_ID of COURSE relation.

STUDENT
ROLLNO NAME ADDRESS COURSE_ID
100 Kanupriya 234, Saraswati Vihar. CS1
101 Rajni Bala 120, Vasant Kunj CS2
102 Arpita Gupta 75, SRM Apartments. CS4

COURSE
COURSE_ID COURSE_NAME DURATION
CS1 MCA 3yrs
CS2 BCA 3yrs
CS3 M.Sc. 2yrs
CS4 B.Sc. 3yrs
CS5 MBA 2yrs

Display the name and other details of all the students along with their course details.

Solution: You may observer that the course details are existing in the COURSE
relation and student names and other details are in the STUDENT relation. Therefore,
you need to join these two relations to extract information. The join attributes in this
case is COURSE_ID in STUDENT and COURSE_ID in the COURSE relation and
equality operator is to be used; therefore, you use natural join operation, which can be
represented as:
STUDENTCOURSE = STUDENT ø COURSE
The output of this natural join will the following relation. Please note that output has
just one COURSE_ID attribute.

STUDENTCOURSE
ROLLNO NAME ADDRESS COURSE_ID COURSE_NAME DURATION
100 Kanupriya 234, Saraswati CS1 MCA 3yrs
Vihar.

3
 101 Rajni Bala 120, Vasant CS2 BCA 3yrs Relational and E-R
Kunj Models
102 Arpita 75, SRM CS4 B.Sc. 3yrs
Gupta Apartments.

There are other types of joins like outer joins. You must refer to further reading for
more details on those operations. They are also explained in later blocks of this
course.

2.4.3 Basic Set Operation

The set operations are the binary operations, i.e., each is applied to two sets or
relations, which should be union compatible. Two relations R (A1, A2,…, An) and S
(B1, B2,…, Bn) are said to be union compatible if they have the same degree n and
domains of the corresponding attributes are also the same i.e.Domain (Ai) = Domain
(Bi) for 1<=i<=n.
Union
If R1 and R2 are two union compatible relations, then R3 = R1 𝖴 R2 is the
relation containing tuples that are either in R1 or in R2 or in both. In other
words, R3 will have tuples such that R3 = {t | t  R1  t  R2}.

Example 9:

R2
R1
A B X Y
A1 B1 A1 B1
A2 B2 A7 B7
A3 B3 A2 B2
A4 B4 A4 B4

R3 = R1 𝖴 R2 is
R3
A B
A1 B1
A2 B2
A3 B3
A4 B4
A7 B7

Note: 1) Union is a commutative operation, i.e,

R 𝖴 S = S 𝖴 R.
2) Union is an associative operation, i.e.
R 𝖴 (S 𝖴 T) = (R 𝖴 S) 𝖴 T.
Intersection

If R1 and R2 are two union compatible functions or relations, then the result of
R3 = R1 ∩ R2 is the relation that includes all tuples that are in both the
relations
In other words, R3 will have tuples such that R3 = {t | t  R1  t  R2}.

3
The Database
Management System
Example 10:
Concepts
R1 R2
X Y A B
A1 B1 A1 B1
A7 B7 A2 B2
A2 B2 A3 B3
A4 B4 A4 B4

R3 = R1 ∩ R2 is

A B
A1 B1
A2 B2
A4 B4
Note: 1) Intersection is a commutative operation, i.e.,
R1 ∩ R2 = R2 ∩ R1.
2) Intersection is an associative operation, i.e.,
R1 ∩ (R2 ∩ R3) = (R1 ∩ R2) ∩
R3
Set Difference
If R1 and R2 are two union compatible relations, then result of R3 =R1– R2 is the
relation that includes only those tuples that are in R1 but not in R2. In other words, R3
will have tuples such that R3 = {t | t  R1  t  R2}.

Example 11:

R1 R2

R1-R2 = X Y A B
A1 B1 A1 B1
A7 B7 A2 B2
A B
A2 B2 A3 B3
A7 B7
A4 B4 A4 B4
R2-R1=

A B
A3 B3

Note: -1) Difference operation is not commutative, i.e.,

R1 – R2 R2 – R1
2) Difference operation is not associative, i. e.,
R1 ─ (R2 – R3)  (R1 – R2) – R3

Please note that a relational query language would be relationally complete, if it

supports five relational algebraic operators – Selection, Projection, Union, Set
Difference and Cartesian Product.

3
 Check Your Progress 1 Relational and E-R
Models

1) A database system is fully relational if it supports a language as powerful as

.

2 Primitive operations are union, difference, product, selection and projection. The
A ∩ B can be computed using as……….

3) Which of the following is not a traditional set operator in relational algebra?

a. Union
b. Intersection
c. Difference
d. Join

4) Consider the relational instances of the relations Suppliers, Parts, proJect and
SPJ relations given below. (Underline represents a key attribute. The SPJ
relation has three Foreign keys: SNO, PNO and JNO.)

S P
SNO SNAME CITY PNO PNAME COLOUR CITY
S1 Smita Delhi P1 Nut Red Delhi
S2 Jim Pune P2 Bolt Blue Pune
S3 Ballav Pune P3 Part1 White Mumbai
S4 Seema Delhi P4 Part2 Blue Delhi
S5 Salim Agra P5 Camera Brown Pune
P6 Part3 Grey Delhi

J SPJ
JNO JNAME CITY SNO PNO JNO QUANTITY
J1 Sorter Pune S1 P1 J1 200
J2 Display Bombay S1 P1 J4 700
J3 OCR Agra S2 P3 J2 400
J4 Console Agra S2 P2 J7 200
J5 RAID Delhi S2 P3 J3 500
J6 EDP Udaipur S3 P3 J5 400
J7 Tape Delhi S3 P4 J3 500
S3 P5 J3 600
S3 P6 J4 800
S4 P6 J2 900
S4 P6 J1 100
S4 P5 J7 200
S5 P5 J5 300
S5 P4 J6 400

Using the in the relations above, which of the following operations and constraints
would be valid:

i. UPDATE Project J7 in J setting CITY to Nagpur.

ii. UPDATE part P5 in P, setting PNO to P4.
iii. UPDATE supplier S5 in S, setting SNO to S8, if the relevant update
rule is RESTRICT.
iv. DELETE supplier S3 in S, if the relevant rule is CASCADE.
v. DELETE part P2 in P, if the relevant delete rule is RESTRICT.
vi. DELETE project J4 in J, if the relevant delete rule is CASCADE.
vii. UPDATE shipment S1-P1-J1 in SPJ, setting SNO to S2. (shipment S1-
P1-J1 means that in SPJ table the value for attributes SNO, PNO and
JNO are S1, P1 and J1 respectively)
3
The Database
Management System
viii.
UPDATE shipment S5-P5-J5 in SPJ, setting JNO to J7.
Concepts ix.
UPDATE shipment S5-P5-J5 in SPJ, setting JNO to J8
x.
INSERT shipment S5-P6-J7 in SPJ.
xi.
INSERT shipment S4-P7-J6 in SPJ
xii.
INSERT shipment S1-P2-jjj (where jjj stands for a default project
number).
……………………………………………………………………………………
……………………………………………………………………………………
……………………………………………………………………………………
5) Find the name of projects in the relations above, to which supplier S1 has
supplied using relational algebra.
…………………………………………………………………………………………….
…………………………………………………………………………………………….
…………………………………………………………………………………….........…
…………………………………………………………………………………………….

2.5 SUMMARY

This unit is an attempt to provide a detailed viewpoint of data base design. The topics
covered in this unit include the relational model including the representation of
relations, operations such as set type operators and relational operators on relations.
The E-R model explained in this unit covers the basic aspects of E-R modeling. E-R
modeling is quite common to database design, therefore, you must attempt as many
problems as possible from the further reading. E-R diagram has also been extended.
However, that is beyond the scope of this unit. You may refer to further readings for
more details on E-R diagrams.

2.6 SOLUTIONS/ ANSWERS

Check Your Progress 1

1. Supplier Number (SNO), Supplier Name (SNAME) and City of the location of the
supplier (CITY). The present instance s of the relation S has 2 tuples.
2. Domain of SNO is the codes that are being assigned to suppliers. The present
coding use first character as S followed by the sequence number of supplier. The
domain of SNAME is strings of characters of some defined length and the domain
of CITY is the set of possible cities in India.
3. The relation will not change; the order of tuples will change.
4. The super key of the relation could be (SNO, SNAME, CITY) or (SNO, SNAME)
or (SNO, CITY) or assuming every supplier has unique name (SNAME, CITY) or
SNO or assuming every supplier has unique name SNAME.
The possible candidate keys are SNO or assuming every supplier has unique name
SNAME
Primary key may be selected as SNO and alternate key would be SNAME

Check Your Progress 2

1. Domain constraint for S will be for CITY which should be checked to be in a

selected list of cities of India. SNO may be checked by a Range constraint and
SNAME may be data type with maximum allowable length of name.

4
2. The Primary key of relation S, P and J are SNO, PNO and JNO respectively. The Relational and E-R
primary key of SPJ is a composite key (SNO, PNO, JNO). As per the primary key Models
constraint none of the attributes of primary key in SPJ, viz. SNO, PNO or JNO
can be null.
3. There are three foreign keys, all of them are in SPJ relation. These are (i) SNO (ii)
PNO (iii) JNO. SNO of SPJ references attribute SNO in S; PNO of SPJ references
attribute PNO in P; and JNO of SPJ references attribute JNO in J. As per foreign
key constraints the values of SNO, PNO and JNO in SPJ can be only as per the
related referenced attributes.
4. The violations are as under:
a. Primary key of S is SNO, it cannot be Null, you may change it to <S4,
Jack, Mumbai>, which will be inserted successfully in S
b. Primary key violation in SPJ as PNO cannot be Null as it is a part of
primary key. There is another violation here, foreign key JNO has a
value J3, which does not exist in JNO attribute of relation J for this
instance, thus, the allowable values for JNO in SPJ is just J1 or J2.
Thus, a correct record insertion in SPJ would be: <’S2’, ‘P3’, ‘J2’,
200>
c. The primary key ‘P2’ already exists and cannot be duplicated, the
correct insertion may be to Insert <’P4, ‘Pencil’, ‘Grey’, ‘Kolkata’>
into P
d. There is no violation.

Check Your Progress 3

1. relational algebra
2. A ─ ( A ─ B)
3. (d) Join
4.
i. Accepted
ii. Rejected (candidate key uniqueness violation)
iii. Rejected (violates RESTRICTED update rule, as SPJ contains tuples
having value S5 in SNO)
iv. Accepted (supplier S3 and all shipments for supplier S3 in the relation
SPJ would be deleted, as the rule is CASCADE)
v. Rejected (violates RESTRICTED delete rule, as SPJ contains tuples
having value P2 in PNO)
vi. Accepted (project J4 and all shipments for project J4 from the relation
SPJ are deleted)
vii. Accepted
viii. Rejected (primary/candidate key uniqueness violation as tuple S5-P5-J7
already exists in relation SPJ)
ix. Rejected (referential integrity violation as there exists no tuple for J8 in
relation J)
x. Accepted
xi. Rejected (referential integrity violation as there exists no tuple for P7 in
relation P)
xii. Rejected (referential integrity violation – the default project number jjj
does not exist in relation J).

5) The answer to this query will require the use of the relational algebraic
operations. This can be found by selection supplies made by S1 in SPJ, then
taking projection of resultant on JNO and joining the resultant to J relation.
Let us show steps:
First find out the supplies made by supplier S1 by selecting those tuples from
SPJ where SNO is S1. The relation operator being:
SPJT =  <SNO = ’S1’> (SPJ)
The resulting temporary relation SPJT will be: 41
The Database
Management System
Concepts SNO PNO JNO QUANTITY
S1 P1 J1 200
S1 P1 J4 700

Next, you may take projection of SPJT on PNO

SPJT2 =  JNO (SPJT)
The resulting temporary relation SPJT will be:

JNO
J1
J4
Next, take natural JOIN this table with J:
RESULT = SPJT2 ø J
The resulting relation RESULT will be:

JNO JNAME CITY

J1 Sorter Pune
J4 Console Agra

You can write it as a single relational algebraic query as:

RESULT = (  JNO ( (SPJ))) ø J
<SNO = ’S1’>