Imperial College London - Databases Relational Algebra

Relational Algebra
Naranker Dulay
n.dulay@imperial.ac.uk
https://www.doc.ic.ac.uk/~nd/databases
The Relational Model
ATTRIBUTES (the columns)

name:type
HEADING title:string year:int length:int genre:string
Gone with
1939 231 Drama
the Wind
BODY TUPLES
Star Wars 1977 124 SF
(the rows)
Wayne’s
1992 95 Comedy
World
Movies RELATION
2
N. Dulay Databases : 02 : Relational Algebra
Relations
Relation = Relation Heading plus Relation Body
Heading = (Unordered) Set of Attributes

Attribute Types are also known as Domains
Attribute = Attribute Name plus Type (normally simple indivisible types).
Why not lists, arrays, graphs, relations?
Body = (Unordered) Set of Tuples
Tuple = Set of attribute values, one for each attribute belonging to attribute type.
Schema for Relation = Name of relation plus heading, for example:
movies(title:string, year:int, length:int, genre:string)
Database = One or more Relations
Schema for Database = Schemas for all relations.
N. Dulay Databases : 02 : Relational Algebra 3

Relations II
In mathematics:
Given sets T1, T2, ... Tn
A relation R is a set of tuples (V1, V2, ... Vn) where Vk ∈ Tk
R is a subset of T1 x T2 ... x Tn.
Tuples are said to be R-related or ‘in R’. e.g. Movies-related or in Movies
n is the degree of the relation, while the number of tuples is the cardinality
Extending to the relational model we have attributed rather than ordered tuples:
R is the set of tuples (A1:T1=V1, ... An:Tn=Vn) where Vk ∈ Tk

Drawing a relation
Important. In the relational model, the ordering of attributes and tuples is

unimportant matter. We can present relations using whatever permutation of
attributes and tuples we like, e.g.:
Movies
year:int genre:string title:string length:int
1977 SF Star Wars 124
Wayne’s
1992 Comedy 95
World
Gone with
1939 Drama 231
the Wind

Drawing a relation II
Relations are not 2-dimensional tables. Although this a convenient way of

presenting them on paper.
A better way to think of them is as a set of n-dimensional values.
For example:
(title:string=StarWars, year:int=1977, length:int=127, genre:string=SF)
is a 4-dimensional movie value.

Relational Algebra
Relational expressions are used construct new relations from other relations.
Operators include:
selection projection product
a x a x
× =
b y a y
c b x
b y
c x
c y
intersection union difference

Applying Set Operations on Relations
Operands must have schemas with an identical set of attributes (name and type).
Before applying union, intersection, or difference we need to ensure that the order of
attributes used for both relations is the same. We can use any ordering we like.
We can use the renaming operator (see later) to change the attribute names of a
relation. This is useful where relations are the same except for one or more attribute
names. For example, an attribute might be called Surname in one relation,
Lastname in another.
For conciseness in the following slides, we shall restrict ourselves to using integer
attributes named a, b, c etc. We’ll omit the type in headings.
Checkout textbooks for more “real-life” relations.

Union R ∪ S
R S R∪S
a b c a b c a b c
1 2 3 7 8 9
4 5 6 2 2 2
7 8 9 3 4 5
1 2 3
Set of tuples in that are in R or in S.

Union R ∪ S
R S R∪S
a b c a b c a b c
1 2 3 7 8 9 1 2 3
4 5 6 2 2 2 4 5 6
7 8 9 3 4 5 7 8 9
1 2 3 2 2 2
3 4 5
Set of tuples in that are in R or in S.
N. Dulay Databases : 02 : Relational Algebra X

Union R ∪ S
R S R∪S
a b c c a b b a c

Difference R − S
R S R−S
a b c a b c a b c
1 2 3 7 8 9
4 5 6 2 2 2
7 8 9 3 4 5
2 2 2 1 2 3
3 4 8
Set of tuples that are in R but not in S

Difference R − S
R S R−S
a b c a b c a b c
1 2 3 7 8 9 4 5 6
4 5 6 2 2 2 3 4 8
7 8 9 3 4 5
2 2 2 1 2 3
3 4 8
Set of tuples in that are in R but not in S

Difference R − S
R S R-S
a b c a b c a b c

R−S≢S−R
R S R−S S−R
a b c a b c a b c a b c
1 2 3 7 8 9
4 5 6 2 2 2
7 8 9 3 4 5
2 2 2 1 2 3
3 4 8
R − S is not equivalent to S − R

R−S≢S−R
R S R−S S−R
a b c a b c a b c a b c
1 2 3 7 8 9 4 5 6 3 4 5
4 5 6 2 2 2 3 4 8
7 8 9 3 4 5
2 2 2 1 2 3
3 4 8
R − S is not equivalent to S − R

Intersection R ∩ S
R S R∩S
a b c a b c a b c
1 2 3 7 8 9
4 5 6 2 2 2
7 8 9 3 4 5
1 2 3
Set of tuples in that are in R and in S.
Intersection can be formulated using difference: R ∩ S = R − (R − S)

R S R∩S
a b c a b c a b c
1 2 3 7 8 9 1 2 3
4 5 6 2 2 2 7 8 9
7 8 9 3 4 5
1 2 3
Set of tuples in that are in R and in S.
Intersection can be formulated using difference: R ∩ S = R − (R − S)

R S R∩S
a b c a b c b a c

Projection πattributes(R)
R πa,c,e(R)
a b c d e f a c e
1 2 3 4 5 6
1 1 1 1 1 1
2 2 2 2 2 2
1 2 3 4 5 8
Projection returns relation with listed attributes only.
Note: no duplicate tuples in the result.

Projection πattributes(R)
R πa,c,e(R)
a b c d e f a c e
1 2 3 4 5 6 1 3 5
1 1 1 1 1 1 1 1 1
2 2 2 2 2 2
2 2 2
1 2 3 4 5 8
Projection returns relation with listed attributes only.
Note: no duplicate tuples in the result.

Selection σcondition(R)
R σa>3 | b<5(R)
a b c a b c
1 2 3
4 5 6
7 8 9
1 5 3
2 9 1
5 8 9
Selection returns relation with all tuples that
2 6 2
satisfy condition.
3 4 5
σ is lowercase sigma (greek alphabet)
R σa>3 | b<5(R)
a b c a b c
1 2 3 1 2 3
4 5 6 4 5 6
7 8 9 7 8 9
1 5 3 5 8 9
2 9 1 3 4 5
5 8 9 Selection returns relation with all tuples that

2 6 2 satisfy condition.
3 4 5 σ is lowercase sigma (greek alphabet)

R σcondition(R)
a b c a b c

Cartesian Product R × S
R S R×S
a b c d e a b c d e
1 2 1 2 3
3 4 4 5 6
7 8 9
Resulting relation has set of all tuples that

can be paired by including a tuple of R and a
tuple from S.
Schema for the result includes the schemas for

both relations. Also known as cross-product.
R S R×S
a b c d e a b c d e
1 2 1 2 3 1 2 1 2 3
3 4 4 5 6 1 2 4 5 6
7 8 9 1 2 7 8 9
3 4 1 2 3
Resulting relation has set of all tuples that
can be paired by including a tuple of R and a
3 4 4 5 6
tuple from S.
Schema for the result includes the schemas for 3 4 7 8 9

both relations. Also known as cross-product.
R S R×S
a b b c d a R.b S.b c d
1 2 1 2 3
3 4 4 5 6
7 8 9
If a R and S have attribute names in

common, then we distinguish them by
prefixing the attribute name with the name
of the relation.

R S R×S
a b b c d a R.b S.b c d
1 2 1 2 3 1 2 1 2 3
3 4 4 5 6 1 2 4 5 6
7 8 9 1 2 7 8 9
3 4 1 2 3
If a R and S have attribute names in
common, then we distinguish them by 3 4 4 5 6
prefixing the attribute name with the name
of the relation. 3 4 7 8 9

Cartesian Product R × S = S × R
R×S S×R
a b c d e c d e a b
1 2 1 2 3 1 2 3 1 2
1 2 4 5 6 1 2 3 3 4
1 2 7 8 9 4 5 6 1 2
3 4 1 2 3 4 5 6 3 4
3 4 4 5 6 7 8 9 1 2
3 4 7 8 9 7 8 9 3 4

Cartesian Product R × S = S × R
R×S S×R
a b c d e c d e a b
1 2 1 2 3 1 2 3 1 2
1 2 4 5 6 1 2 3 3 4
1 2 7 8 9 4 5 6 1 2
3 4 1 2 3 4 5 6 3 4
3 4 4 5 6 7 8 9 1 2
3 4 7 8 9 7 8 9 3 4

Natural Join R ⋈ S
R S R⋈S
a b b c d a b c d
1 2 1 2 3
1 3 2 5 6
3 4 4 8 9
4 9 9
Resulting relation has all tuples that can be “joined” using all matching attributes
of R and S.
Resulting schema is the set of both schemas.

R S R⋈S
a b b c d a b c d
1 2 1 2 3 1 2 5 6
1 3 2 5 6 3 4 8 9
3 4 4 8 9 3 4 9 9
4 9 9
Resulting relation has all tuples that can be “joined” using all matching attributes
of R and S.
Resulting schema is the set of both schemas.

R S R⋈S
a b c b c d a b c d
1 2 3 2 3 3
6 7 8 2 3 6
9 7 8 4 8 9
4 2 7 7 8 3
1 7 8
Must only pair tuples if values of each common attribute are equal.

R S R⋈S
a b c b c d a b c d
1 2 3 2 3 3 1 2 3 3
6 7 8 2 3 6 1 2 3 6
9 7 8 4 8 9 6 7 8 3
4 2 7 7 8 3 9 7 8 3
1 7 8 1 7 8 3
Must only pair tuples if values of each common attribute are equal.

R S R⋈S
a b c a c d e a b c d e

Natural Join
Natural join can be expressed in terms of product, selection and projection:
R⋈S= πL (σC(R × S))

where L is union of all attributes of R and S i.e. attr(R) ∪ attr(S)
C is R.a1=S.a1 & R.a2 = S.a2 & ... & R.an = S.an

where ak are the attributes common to R and S, ak ∈attr(R) ∩ attr(S)
Natural join is associative: (R ⋈ S) ⋈ T = R ⋈ (S ⋈ T)
If there is no common attribute then: R ⋈ S = R × S i.e. we don’t apply selection

and projection.

Renaming ρnew /old , new /old , ..., new /old (R)
1 1 2 2 n n
R ρx/a, y/d(R)
a b c d
1 2 5 6 1 2 5 6
3 4 8 9 3 4 8 9
3 4 9 9 3 4 9 9
Changes the names of one or more attributes. ρ is lowercase rho

Note: the values and types of the relation are not affected.
Renaming ρnew /old , new /old , ..., new /old (R)
1 1 2 2 n n
R ρx/a, y/d(R)
a b c d x b c y
1 2 5 6 1 2 5 6
3 4 8 9 3 4 8 9
3 4 9 9 3 4 9 9
Changes the names of one or more attributes.
Note: the values and types of the relation are not affected.
Relational Expressions
Relational expressions are used to construct new relations from other relations in:
Retrieval - to define data to be retrieved.
Updates - to define data to be inserted, changed, or deleted.
Constraints - to define constraints that the database must satisfy.
Derived relations (views) - to define one relation in terms of others, without

constructing it.
Concurrency - to define the data to use for concurrent actions.
Access control - to define the data over which a permission should be granted.
A relational database language is said to be relationally complete if it is at least as

expressive as Relational Algebra.

Exercise 1
Consider the database schema (types omitted):
employee(personname, street, city)

worksfor(personname, companyname, salary)
company(companyname, city)
Give relational expressions for each of the following:
1. Names of all employees who work for Barclays.

2. Names of Barclays employees and the city where they live.
3. Names, street and city of all Barclays employees who earn more than £1M.
4. What will πpersonname(employee ⋈ worksfor ⋈

company) return?
Exercise 1

1. Names of all employees who work for Barclays.
πpersonname(σcompanyname=”Barclays”(worksfor))

Exercise 1

2. Names of Barclays employees and city where they live.
πpersonname, city(
(σcompanyname=”Barclays”(workfor) ⋈ employee))

Exercise 1

3. Names, street and city of all Barclays employees who earn more than £1M.
πpersonname, street, city(

σcompanyname=”Barclays” & salary>1000000(worksfor)
⋈ employee)

Exercise 1

4. What will
πpersonname (employee ⋈ worksfor ⋈ company)

return?
Names of all employees who live in the same city as the company they work for.

Exercise 2
Consider the database schema (types omitted):
supplier(sname, address)
part(pname, colour)
catalog(sname, pname, cost)
Give relational expressions for each of the following:
1. Names of all suppliers who supply a red part.

2. Names of all suppliers who supply a red part or are at 18 Queensgate
3. Names of parts supplied by at least two different suppliers.
Assume: C1 and C2 are temporary copies of the catalog relation.
4. Pairs of suppliers such that the first suppliers charges more for some part than the
second supplier. Assume: C1 and C2 are temporary copies of the catalog relation.

Exercise 2
part(pname, colour)
1. Names of all suppliers who supply a red part
πsname(σcolour=”red”(part) ⋈ catalog)

Exercise 2
part(pname, colour)
2. Names of all suppliers who supply a red part or are at 180 Queensgate
πsname(σcolour=”red”(part) ⋈ catalog) ∪
πsname(σaddress=”180 Queensgate”(supplier))

Exercise 2
part(pname, colour)
3. Names of parts supplied by at least two different suppliers.
Assume: C1 and C2 are copies of the catalog relation.
πC1.pname(
σC1.pname=C2.pname & C1.sname≠C2.sname(C1×C2))

Exercise 2
part(pname, colour)
4. Pairs of suppliers such that the first supplier charges more for some part than the
second supplier. Assume: C1 and C2 are temporary copies of the catalog relation.
πC1.sname, C2.sname(
σC1.pname=C2.pname & C1.sname≠C2.sname &
C1.cost>C2.cost(C1 × C2))

Imperial College London - Databases Relational Algebra

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Imperial College London - Databases Relational Algebra

Uploaded by

Copyright:

Available Formats

Relational Algebra

ATTRIBUTES (the columns)

HEADING title:string year:int length:int genre:string

Relation = Relation Heading plus Relation Body

Heading = (Unordered) Set of Attributes

Schema for Relation = Name of relation plus heading, for example:

movies(title:string, year:int, length:int, genre:string)

Database = One or more Relations

Schema for Database = Schemas for all relations.

N. Dulay Databases : 02 : Relational Algebra 3

Given sets T1, T2, ... Tn

A relation R is a set of tuples (V1, V2, ... Vn) where Vk ∈ Tk

R is a subset of T1 x T2 ... x Tn.

Tuples are said to be R-related or ‘in R’. e.g. Movies-related or in Movies

R is the set of tuples (A1:T1=V1, ... An:Tn=Vn) where Vk ∈ Tk

N. Dulay Databases : 02 : Relational Algebra 4

Important. In the relational model, the ordering of attributes and tuples is

year:int genre:string title:string length:int

1977 SF Star Wars 124

N. Dulay Databases : 02 : Relational Algebra 5

Relations are not 2-dimensional tables. Although this a convenient way of

A better way to think of them is as a set of n-dimensional values.

(title:string=StarWars, year:int=1977, length:int=127, genre:string=SF)

is a 4-dimensional movie value.

N. Dulay Databases : 02 : Relational Algebra 6

selection projection product

N. Dulay Databases : 02 : Relational Algebra 7

N. Dulay Databases : 02 : Relational Algebra 8

Set of tuples in that are in R or in S.

N. Dulay Databases : 02 : Relational Algebra 9

Set of tuples in that are in R or in S.

N. Dulay Databases : 02 : Relational Algebra X

N. Dulay Databases : 02 : Relational Algebra 10

N. Dulay Databases : 02 : Relational Algebra 11

N. Dulay Databases : 02 : Relational Algebra X

N. Dulay Databases : 02 : Relational Algebra 12

N. Dulay Databases : 02 : Relational Algebra 13

N. Dulay Databases : 02 : Relational Algebra X

Set of tuples in that are in R and in S.

Intersection can be formulated using diﬀerence: R ∩ S = R − (R − S)

N. Dulay Databases : 02 : Relational Algebra 14

Set of tuples in that are in R and in S.

Intersection can be formulated using diﬀerence: R ∩ S = R − (R − S)

N. Dulay Databases : 02 : Relational Algebra X

N. Dulay Databases : 02 : Relational Algebra 15

Projection returns relation with listed attributes only.

Note: no duplicate tuples in the result.

N. Dulay Databases : 02 : Relational Algebra 16

Projection returns relation with listed attributes only.

Note: no duplicate tuples in the result.

N. Dulay Databases : 02 : Relational Algebra X

5 8 9 Selection returns relation with all tuples that

3 4 5 σ is lowercase sigma (greek alphabet)

N. Dulay Databases : 02 : Relational Algebra X

N. Dulay Databases : 02 : Relational Algebra 18

Resulting relation has set of all tuples that

Schema for the result includes the schemas for

Schema for the result includes the schemas for 3 4 7 8 9

If a R and S have attribute names in

N. Dulay Databases : 02 : Relational Algebra 20

N. Dulay Databases : 02 : Relational Algebra X

N. Dulay Databases : 02 : Relational Algebra 21

N. Dulay Databases : 02 : Relational Algebra X