Relational Algebra

VICTORIA UNIVERSITY OF WELLINGTON
Te Whare Wananga o te Upoko o te Ika a Maui
Relational Algebra
Lecturer Dr Pavle Mogin
COMP302
Database Systems
Plan for Relational Algebra
Relational database operations

Set theoretic operations
Additional operations
Reading:
Chapters 6 of the textbook
COMP302 Database Systems 2007
Relational Algebra 1
Relational Algebra
One of the relational data model languages
The others are Tuple Calculus, Domain Calculus,
SQL
Important for query optimization
A SQL query is transformed into a relational algebra
equivalent, optimized and executed
A set of operations to manipulate (query and update)

a relational database
Operations are applied onto relations
The result is a new relation
Relational Algebra Operations

Three types of operations:
Relational database operations:
select, project, and join
Set theoretic operations:

union, intersect, set difference, Cartesian product
Additional relational operations:

aggregate operations, grouping, and outer join
Select operation
It is used to select such a subset of tuples from a
relation that satisfy given condition
Notation: C (r (N ))
Condition C is a Boolean expression on attributes of
R in N(R, C)
Boolean expression is made up of clauses of the form
A a or A B, where adom (A ), { =, <, >, <, >, },
and A, B R, ( is equivalent to <>)
Clauses can be connected by Boolean operators , ,
to form new clauses
A Sample Relational Database

Student
LName FName StudId
Major
Smith
Susan
131313 Comp
Bond
James
007007 Math
Smith
Susan
555555 Comp
Cecil
John
010101 Math
Course
PName CourId
Points
Dept
DB Sys C302
15
Comp
SofEng C301
15
Comp
DisMat
M214
22
Math
Pr&Sys C201
22
Comp
Grades
StudId
CourId
Grade
007007
C302
A+
555555
C302
007007
C301
007007
M214
A+
131313
C201
B-
555555
C201
131313
C302
007007
C201
010101
C201
Select Operation (continued)

Relation C r(N) contains those tuples of r(N) that
evaluate true for C, hence:
degree(C r (N )) = degree(r (N )) (number of attributes)
C r (N ) r (N ) and | C r (N ) | < |r (N ) | (number of tuples)
e.g.
Student1
StudId = 007007 r (Student ) = LName FName StudId Major
Bond
James
007007 Math
Student2
LName FName StudId
FName = 'Susan' r (Student ) =

Major
Smith
Susan
131313 Comp
Smith
Susan
555555 Comp
Project Operation
Project operation produces a new relation by
choosing some of the columns from the starting
relation, and dropping all the others
Notation:
ALr (N ),
where AL is a attribute sublist from R in N (R, C)
If AL = (Al,, Ak), then ALr (R ) = r (N )[Al,, Ak]
Student1
e.g.
<LName, FName>r (Student ) =

Lname
FName
Smith
Susan
Bond
James
Cecil
John
Project Operation (continued)

Properties of project operation:
Degree(ALr (R )) = |AL | < |R | (number of attributes)
|ALr (R )| < |r (R )| (number of tuples)
Combining select and project operations
AL (C r (R ))
e.g.
(FName, LName ) (StudId = 007007 r (Student )) =
Student3
FName
LName
James
Bond
Join Operation
Join operation joins two relations by merging those
tuples from two relations that satisfy a given condition
The condition is defined on attributes belonging to relations to
be joined
Theta, equi, and natural join operations

Theta, equi, and natural join are collectively called
INNER joins
In each of inner joins, tuples with null valued join
attributes do not appear in the result
Theta Join Operation

Notation:
r (N) = r (N1) JC r (N2)
r(N ) is the result of joining r(N1) over N1(R1, C1) with r(N2)
over N2(R2, C2)
Join condition JC = jc1 jcn
jci = A B, AR1, BR2,
{=, , <, >, , },
Dom(N1, A ) Dom (N2, B ),
Range(N1, A ) Dom (N2, B )
R1 = {A1,, An}, R2 = {B1,, Bn}, R = {A1,, An, B1,, Bn}

degree(r (N )) = degree(r (N1)) + degree(r (N2))
| r (R) | < | r (R1) | | r (R2) |
Equijoin Operation
A special case of the theta join, when {=}

Notation: r (N ) = r (N1) X = Y r (N2), X R1, Y R2
e.g.
Let X = Student.StudId and Y = Grades.StudId and
r (Stud_Grades) = r (Student ) X = Y r (Grades ),
then
Equijoin Operation (continued)

Superfluous
column
Stud_Grades
StudId
StudId
Susan
131313
131313
Comp
C201
B-
Smith
Susan
131313
131313
Comp
C302
Bond
James
007007
007007
Math
C302
A+
Bond
James
007007
007007
Math
C301
Bond
James
007007
007007
Math
M214
A+
Bond
James
007007
007007
Math
C201
Smith
Susan
555555
555555
Comp
C201
Smith
Susan
555555
555555
Comp
C302
Cecil
John
010101
010101
Math
C201
LName
FName
Smith
Major
CourId
Grade
Natural Join Operation

A special case of an equijoin operation, when
attributes belonging to different relation schemas in
the equijoin condition have the same name (N1.X =
N2.X )
Notation: r (N) = r (N1) * r (N2)
Formal definition:
r (N1) * r (N2) = { t [R1 R2] | t [R1]r (N1) t [R2]r (N2)}
degree(r (N) = degree(r (N1) + degree(r (N2) |X |) (no of attr)
0 < |r (N1) * r (N2)| < |r (N1)| |r (N1)| (no of tuples)
Set Theoretic Operations

Union, Intersect, Difference, Cartesian product
r(N) = r(N1) r (N1),
where R1 = (A1,, An), R2 = (B1,, Bn) are lists of attributes, and
{ , , \, } ( \ or -)
For union, intersect and difference attribute sets R1 and R2 have
to be union compatible:
| R1| = | R2|,
(i{1,, n })(Dom (N1, Ai ) = Dom (N2, Bi)), and
(i{1,, n })(Range (N1, Ai ) = Range (N2, Bi ))
Cartesian product
R = R1 R2, degree(r(N1) r(N2)) = degree(r(N1)) +
degree(r(N2 )),
|r (N1) r (N2)| = |r (N1) | |r (N2)|
Question For You
Consider the following relations

r (AB )
r (BC )
How many tuples will contain the Cartesian product

r(AB ) r(BC )?
a) 3
b) 6
c) 9
Additional relational operations

Relational database and set theoretic operations are not
sufficient to perform some common database queries
To enhance the power of relational algebra, there are some new
operations introduced:
Aggregate functions (SUM, AVERAGE, MAX, MIN, COUNT),
Grouping, and
Outer join
Aggregate functions are defined on numeric attributes and

applied on all relation tuples
Relation is partitioned by means of grouping attribute values,
and the defined aggregate functions are computed for the tuples
of each group
Outer join extends the join operation to cope with null values
Aggregate Functions and Grouping

Notation:
<grouping attributes>Y<(function, attribute) list>(r (N ))
where:
<grouping attributes> is a list of attributes from R,
Y (pronounced as "script F") is the symbol used to denote
aggregate operation, and
<(function, attribute)_list> is a list of pairs (aggregate function
from {SUM, AVERAGE, COUNT, MIN, MAX }, attribute from
R)
The resulting relation has columns for grouping

attributes, and one column with the name of the form
FUNCTION_ATTRIBUTE for each (function, attribute)
pair
Outer Join
Introduced to insert those tuples that don't match, or
contain null values for join attributes into join relation
Notation:
LEFT, RIGHT, and FULL outer join
Example:
r (AB )
A
B = B
r (BC )
r (ABBC )
Summary
Relational algebra:
Relational database operations,
Additional operations
Relational database operations

select (chooses tuples), project (restricts tuples), join
(concatenates matching tuples)

Union, Intersect, Set difference, Cartesian product
Additional relational operations

Aggregate functions COUNT, SUM, AVERAGE,, Grouping,
and Outer join
Next topic
Structured Query Language

Reading:
Chapter 8 of the textbook

Relational Algebra

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Relational Algebra

Caricato da

Copyright:

Formati disponibili

VICTORIA UNIVERSITY OF WELLINGTON

Te Whare Wananga o te Upoko o te Ika a Maui

Plan for Relational Algebra

Relational database operations

COMP302 Database Systems 2007

A set of operations to manipulate (query and update)