Database Systems

Session 5 – Main Theme

Relational Algebra, Relational Calculus, and SQL

Dr. Jean-Claude Franchitti

New York University

Computer Science Department
Courant Institute of Mathematical Sciences

Presentation material partially based on textbook slides

Fundamentals of Database Systems (6th Edition)
by Ramez Elmasri and Shamkant Navathe
Slides copyright © 2011 and on slides produced by Zvi
Kedem copyight © 2014

1
Agenda

1 Session Overview

2 Relational Algebra and Relational Calculus

3 Relational Algebra Using SQL Syntax

5 Summary and Conclusion

2
Session Agenda

 Session Overview
 Relational Algebra and Relational Calculus
 Relational Algebra Using SQL Syntax
 Summary & Conclusion

3
What is the class about?

 Course description and syllabus:

» http://www.nyu.edu/classes/jcf/CSCI-GA.2433-001
» http://cs.nyu.edu/courses/fall11/CSCI-GA.2433-001/

 Textbooks:
» Fundamentals of Database Systems (6th Edition)
Ramez Elmasri and Shamkant Navathe
Addition Wesley
ISBN-10: 0-1360-8620-9, ISBN-13: 978-0136086208 6th Edition (04/10)

4
Icons / Metaphors

Information

Common Realization

Knowledge/Competency Pattern

Governance

Alignment

Solution Approach

55
Agenda

1 Session Overview

2 Relational Algebra and Relational Calculus

3 Relational Algebra Using SQL Syntax

5 Summary and Conclusion

6
Agenda

 Unary Relational Operations: SELECT and

PROJECT
 Relational Algebra Operations from Set Theory
 Binary Relational Operations: JOIN and
DIVISION
 Additional Relational Operations
 Examples of Queries in Relational Algebra
 The Tuple Relational Calculus
 The Domain Relational Calculus

7
The Relational Algebra and Relational Calculus

 Relational algebra
 Basic set of operations for the relational model
 Relational algebra expression
 Sequence of relational algebra operations
 Relational calculus
 Higher-level declarative language for specifying
relational queries

8
Unary Relational Operations: SELECT and PROJECT (1/3)

 The SELECT Operation

 Subset of the tuples from a relation that
satisfies a selection condition:

• Boolean expression contains clauses of the form

<attribute name> <comparison op> <constant
value>
• or
• <attribute name> <comparison op> <attribute
name>

9
Unary Relational Operations: SELECT and PROJECT (2/3)

 Example:

 <selection condition> applied independently

to each individual tuple t in R
 If condition evaluates to TRUE, tuple selected
 Boolean conditions AND, OR, and NOT
 Unary
 Applied to a single relation

10
Unary Relational Operations: SELECT and PROJECT (3/3)

 Selectivity
 Fraction of tuples selected by a selection
condition
 SELECT operation commutative
 Cascade SELECT operations into a single
operation with AND condition

11
The PROJECT Operation

 Selects columns from table and discards

the other columns:

 Degree
 Number of attributes in <attribute list>
 Duplicate elimination
 Result of PROJECT operation is a set of
distinct tuples

12
Sequences of Operations and the RENAME Operation

 In-line expression:

 Sequence of operations:

 Rename attributes in intermediate results

 RENAME operation

13
Relational Algebra Operations from Set Theory (1/2)

 UNION, INTERSECTION, and MINUS

 Merge the elements of two sets in various ways
 Binary operations
 Relations must have the same type of tuples
 UNION
RUS
 Includes all tuples that are either in R or in S or
in both R and S
 Duplicate tuples eliminated

14
Relational Algebra Operations from Set Theory (2/2)

 INTERSECTION
R∩S
 Includes all tuples that are in both R and S
 SET DIFFERENCE (or MINUS)
R–S
 Includes all tuples that are in R but not in S

15
The CARTESIAN PRODUCT (CROSS PRODUCT) Operation

 CARTESIAN PRODUCT
 CROSS PRODUCT or CROSS JOIN
 Denoted by ×
 Binary set operation
 Relations do not have to be union compatible
 Useful when followed by a selection that
matches values of attributes

16
Binary Relational Operations: JOIN and DIVISION (1/2)

 The JOIN Operation

 Denoted by
 Combine related tuples from two relations into
single “longer” tuples
 General join condition of the form <condition>
AND <condition> AND...AND <condition>
 Example:

17
Binary Relational Operations: JOIN and DIVISION (2/2)

 THETA JOIN
 Each <condition> of the form Ai θ Bj
 Ai is an attribute of R
 Bj is an attribute of S
 Ai and Bj have the same domain
 θ (theta) is one of the comparison operators:
• {=, <, ≤, >, ≥, ≠}

18
Variations of JOIN: The EQUIJOIN and NATURAL JOIN (1/2)

 EQUIJOIN
 Only = comparison operator used
 Always have one or more pairs of attributes
that have identical values in every tuple
 NATURAL JOIN
 Denoted by *
 Removes second (superfluous) attribute in an
EQUIJOIN condition

19
Variations of JOIN: The EQUIJOIN and NATURAL JOIN (2/2)

 Join selectivity
 Expected size of join result divided by the
maximum size nR * nS
 Inner joins
 Type of match and combine operation
 Defined formally as a combination of
CARTESIAN PRODUCT and SELECTION

20
A Complete Set of Relational Algebra Operations

 Set of relational algebra operations {σ, π,

U, ρ, –, ×} is a complete set
 Any relational algebra operation can be
expressed as a sequence of operations from
this set

21
The DIVISION Operation

 Denoted by ÷
 Example: retrieve the names of employees
who work on all the projects that ‘John
Smith’ works on
 Apply to relations R(Z) ÷ S(X)
 Attributes of R are a subset of the attributes of
S

22
Operations of Relational Algebra (1/2)

23
Operations of Relational Algebra (2/2)

24
Notation for Query Trees

 Query tree
 Represents the input relations of query as leaf
nodes of the tree
 Represents the relational algebra operations
as internal nodes

25
Sample Query Tree for Relational Algebra Expression

26
Additional Relational Operations (1/2)

 Generalized projection
 Allows functions of attributes to be included in
the projection list

 Aggregate functions and grouping

 Common functions applied to collections of
numeric values
 Include SUM, AVERAGE, MAXIMUM, and
MINIMUM

27
Additional Relational Operations (2/2)

 Group tuples by the value of some of their

attributes
 Apply aggregate function independently to
each group

28
Sample Aggregate Function Operation

29
Recursive Closure Operations

 Operation applied to a recursive

relationship between tuples of same type

30
OUTER JOIN Operations

 Outer joins
 Keep all tuples in R, or all those in S, or all
those in both relations regardless of whether or
not they have matching tuples in the other
relation
 Types
• LEFT OUTER JOIN, RIGHT OUTER JOIN, FULL
OUTER JOIN
 Example:

31
The OUTER UNION Operation

 Take union of tuples from two relations that

have some common attributes
 Not union (type) compatible
 Partially compatible
 All tuples from both relations included in the
result
 Tuples with the same value combination will
appear only once

32
Examples of Queries in Relational Algebra (1/3)

33
Examples of Queries in Relational Algebra (2/3)

34
Examples of Queries in Relational Algebra (3/3)

35
The Tuple Relational Calculus

 Declarative expression
 Specify a retrieval request
 Non-procedural language
 Any retrieval that can be specified in basic
relational algebra
 Can also be specified in relational calculus

36
Tuple Variables and Range Relations

 Tuple variables
 Ranges over a particular database relation
 Satisfy COND(t):
 Specify:
 Range relation R of t
 Select particular combinations of tuples
 Set of attributes to be retrieved (requested
attributes)

37
Expressions and Formulas in Tuple Relational Calculus

 General expression of tuple relational

calculus is of the form:

 Truth value of an atom

 Evaluates to either TRUE or FALSE for a
specific combination of tuples
 Formula (Boolean condition)
 Made up of one or more atoms connected via
logical operators AND, OR, and NOT

38
Existential and Universal Quantifiers

 Universal quantifier (inverted “A”)

 Existential quantifier (mirrored “E”)
 Define a tuple variable in a formula as free
or bound

39
Sample Queries in Tuple Relational Calculus

40
Notation for Query Graphs

41
Transforming the Universal and Existential Quantifiers

 Transform one type of quantifier into other

with negation (preceded by NOT)
 AND and OR replace one another
 Negated formula becomes un-negated
 Un-negated formula becomes negated

42
Using the Universal Quantifier in Queries

43
Safe Expressions

 Guaranteed to yield a finite number of

tuples as its result
 Otherwise expression is called unsafe
 Expression is safe
 If all values in its result are from the domain of
the expression

44
The Domain Relational Calculus (1/2)

 Differs from tuple calculus in type of

variables used in formulas
 Variables range over single values from
domains of attributes
 Formula is made up of atoms
 Evaluate to either TRUE or FALSE for a
specific set of values
• Called the truth values of the atoms

45
The Domain Relational Calculus (2/2)

 QBE language
 Based on domain relational calculus

46
Summary

 Formal languages for relational model of

data:
 Relational algebra: operations, unary and
binary operators
 Some queries cannot be stated with basic
relational algebra operations
• But are important for practical use
 Relational calculus
 Based predicate calculus

47
Agenda

1 Session Overview

2 Relational Algebra and Relational Calculus

3 Relational Algebra Using SQL Syntax

4 Summary and Conclusion

48
Agenda

 Relational Algebra and SQL

 Basic Syntax Comparison
 Sets and Operations on Relations
 Relations in Relational Algebra
 Empty Relations
 Relational Algebra vs. Full SQL
 Operations on Relations
» Projection
» Selection
» Cartesian Product
» Union
» Difference
» Intersection
 From Relational Algebra to Queries (with Examples)
 Microsoft Access Case Study
 Pure Relational Algebra
49
Relational Algebra And SQL

 SQL is based on relational algebra with many extensions

» Some necessary
» Some unnecessary
 “Pure” relational algebra, use mathematical notation with Greek
letters
 It is covered here using SQL syntax; that is this unit covers
relational algebra, but it looks like SQL
And will be really valid SQL
 Pure relational algebra is used in research, scientific papers, and
some textbooks
 So it is good to know it, and material is provided at the end of this
unit material from which one can learn it
 But in anything practical, including commercial systems, you will be
using SQL

50
Key Differences Between SQL And “Pure” Relational Algebra

 SQL data model is a multiset not a set; still rows in tables (we
sometimes continue calling relations)
» Still no order among rows: no such thing as 1st row
» We can (if we want to) count how many times a particular row appears
in the table
» We can remove/not remove duplicates as we specify (most of the time)
» There are some operators that specifically pay attention to duplicates
» We must know whether duplicates are removed (and how) for each
SQL operation; luckily, easy
 Many redundant operators (relational algebra had only one:
intersection)
 SQL provides statistical operators, such as AVG (average)
» Can be performed on subsets of rows; e.g. average salary per
company branch

51
Key Differences Between Relational Algebra And SQL

 Every domain is “enhanced” with a special

element: NULL
» Very strange semantics for handling these
elements
 “Pretty printing” of output: sorting, and
similar
 Operations for
» Inserting
» Deleting
» Changing/updating (sometimes not easily
reducible to deleting and inserting)
52
Basic Syntax Comparison (1/2)

Relational Algebra SQL

p a, b SELECT a, b

s (d > e)  (f =g ) WHERE d > e AND f = g

pq FROM p, q

p a, b s (d > e)  (f =g ) (p  q) SELECT a, b
FROM p, q
WHERE d > e AND f = g;
{must always have SELECT even if all
attributes are kept, can be written as:
SELECT *}
renaming AS {or blank space}

p := result INSERT INTO p

result {assuming p was empty}
p a, b (p) (assume a, b are the only SELECT *
attributes) FROM p;
53
Basic Syntax Comparison (2/2)

Relational Algebra SQL

pq SELECT *
FROM p
UNION
SELECT *
FROM q
p−q SELECT *
FROM p
EXCEPT
SELECT *
FROM q
Sometimes, instead, we have DELETE
FROM
pq SELECT *
FROM p
INTERSECT
SELECT *
FROM q

54
Sets And Operations On Them

 If A, B, and C are sets, then we have the operations

  Union, A  B = { x  x  A  x  B }
  Intersection, A  B = { x  x  A  x  B }
  Difference, A  B = { x  x  A  x  B }
  Cartesian product, A  B = { (x,y)  x  A  y  B }, A
 B  C = { (x,y,z)  x  A  y  B  z  B }, etc.
 The above operations form an algebra, that is you can
perform operations on results of operations, such as (A
 B)  (C  A)
So you can write expressions and not just programs!

55
Relations in Relational Algebra

 Relations are sets of tuples, which we will also call

rows, drawn from some domains
 Relational algebra deals with relations (which look like
tables with fixed number of columns and varying
number of rows)
 We assume that each domain is linearly ordered, so for
each x and y from the domain, one of the following
holds
» x<y
» x=y
» x<y
 Frequently, such comparisons will be meaningful even if
x and y are drawn from different columns
» For example, one column deals with income and another with
expenditure: we may want to compare them
56
Reminder: Relations in Relational Algebra

 The order of rows and whether a row appears once or

many times does not matter
 The order of columns matters, but as our columns will
always be labeled, we will be able to reconstruct the
order even if the columns are permuted.
 The following two relations are equal:

R A B R B A
1 10 20 2
2 20 10 1
20 2
20 2

57
Many Empty Relations

 In set theory, there is only one empty set

 For us, it is more convenient to think that for each
relation schema, that for specific choice of column
names and domains, there is a different empty relation
 And of, course, two empty relations with different
number of columns must be different
 So for instance the two relations below are different

 The above needs to be stated more precisely to be

“completely correct,” but as this will be intuitively clear,
we do not need to worry about this too much

58
Relational Algebra Versus Full SQL

 Relational algebra is restricted to querying the database

 Does not have support for
» Primary keys
» Foreign keys
» Inserting data
» Deleting data
» Updating data
» Indexing
» Recovery
» Concurrency
» Security
» …
 Does not care about efficiency, only about specifications
of what is needed

59
Operations on relations

 There are several fundamental operations on

relations
 We will describe them in turn:
» Projection
» Selection
» Cartesian product
» Union
» Difference
» Intersection (technically not fundamental)
 The very important property: Any operation on
relations produces a relation
 This is why we call this an algebra
60
Projection: Choice Of Columns

R A B C D
1 10 100 1000
1 20 100 1000
1 20 200 1000

 SQL statement B A D
SELECT B, A, D 10 1 1000
20 1 1000
FROM R
20 1 1000

 We could have removed the duplicate

row, but did not have to
61
Selection: Choice Of Rows
R A B C D
5 5 7 4
5 6 5 7
4 5 4 4
5 5 5 5
4 6 5 3
4 4 3 4
 SQL statement:
SELECT * (this means all columns) 4 4 4 5
FROM R 4 6 4 6
WHERE A <= C AND D = 4; (this is a predicate, i.e., condition)

A B C D
5 5 7 4
4 5 4 4
62
Selection

 In general, the condition (predicate) can be

specified by a Boolean formula with
NOT, AND, OR on atomic conditions, where a
condition is:
» a comparison between two column names,
» a comparison between a column name and a
constant
» Technically, a constant should be put in quotes
» Even a number, such as 4, perhaps should be put in
quotes, as ‘4’, so that it is distinguished from a
column name, but as we will never use numbers for
column names, this not necessary

63
Cartesian Product
RA B S C B D
1 10 40 10 10
2 10 50 20 10
2 20
 SQL statement
SELECT A, R.B, C, S.B, D
FROM R, S; (comma stands for Cartesian product)
A R.B C S.B D
1 10 40 10 10
1 10 50 20 10
2 10 40 10 10
2 10 50 20 10
2 20 40 10 10
2 20 50 20 10
64
A Typical Use Of Cartesian Product

R Size Room# S ID# Room# YOB

140 1010 40 1010 1982
150 1020 50 1020 1985
140 1030

 SQL statement:
SELECT ID#, R.Room#, Size
FROM R, S
WHERE R.Room# = S.Room#;
ID# R.Room# Size
40 1010 140
50 1020 150
65
A Typical Use Of Cartesian Product

 After the Cartesian product, we got

Size R.Room# ID# S.Room YOB
#
140 1010 40 1010 1982
140 1010 50 1020 1985
150 1020 40 1010 1982
150 1020 50 1020 1985
140 1030 40 1010 1982

140 1030 50 1020 1985

 This allowed us to correlate the information from

the two original tables by examining each tuple
in turn
66
A Typical Use Of Cartesian Product

 This example showed how to correlate information from

two tables
» The first table had information about rooms and their sizes
» The second table had information about employees including
the rooms they sit in
» The resulting table allows us to find out what are the sizes of
the rooms the employees sit in
 We had to specify R.Room# or S.Room#, even though
they happen to be equal due to the specific equality
condition
 We could, as we will see later, rename a column, to get
Room#
ID# Room# Size
40 1010 140
50 1020 150
67
Union

R A B S A B
1 10 1 10
2 20 3 20

 SQL statement
(SELECT * A B
FROM R) 1 10
UNION
2 20
(SELECT *
FROM S); 3 20

 Note: We happened to choose to remove duplicate

rows
 Note: we could not just write R UNION S (syntax quirk)
68
Union Compatibility

 We require same -arity (number of columns),

otherwise the result is not a relation
 Also, the operation “probably” should make
sense, that is the values in corresponding
columns should be drawn from the same
domains
 Actually, best to assume that the column names
are the same and that is what we will do from
now on
 We refer to these as union compatibility of
relations
 Sometimes, just the term compatibility is used
69
Difference

R A B S A B
1 10 1 10
2 20 3 20
 SQL statement
(SELECT *
FROM R)
MINUS A B
(SELECT *
2 20
FROM S);
 Union compatibility required
 EXCEPT is a synonym for MINUS

70
Intersection

R A B S A B
1 10 1 10
2 20 3 20

 SQL statement
(SELECT *
FROM R) A B
INTERSECT 1 10
(SELECT *
FROM S);
 Union compatibility required
 Can be computed using differences only: R – (R – S)

71
From Relational Algebra to Queries

 These operations allow us to define a large

number of interesting queries for relational
databases.
 In order to be able to formulate our examples,
we will assume standard programming
language type of operations:
» Assignment of an expression to a new variable;
In our case assignment of a relational expression to
a relational variable.
» Renaming of a relations, to use another name to
denote it
» Renaming of a column, to use another name to
denote it
72
A Small Example

 The example consists of 3 relations:

 Person(Name,Sex,Age)
 This relation, whose primary key is Name, gives information about the
human’s sex and age
 Birth(Parent,Child)
 This relation, whose primary key is the pair Parent,Child, with both being
foreign keys referring to Person gives information about who is a parent of
whom. (Both mother and father would be generally listed)
 Marriage(Husband,Wife, Age) or
 Marriage(Husband,Wife, Age)
 This relation listing current marriages only, requires choosing which spouse
will serve as primary key. For our exercise, it does not matter what the
choice is. Both Husband and Wife are foreign keys referring to Person. Age
specifies how long the marriage has lasted.
 For each attribute above, we will frequently use its first letter to refer to
it, to save space in the slides, unless it creates an ambiguity
 Some ages do not make sense, but this is fine for our example

73
Relational Implementation

 Two options for selecting the primary key of Marriage

 The design is not necessarily good, but nice and simple
for learning relational algebra
Marriage Person Birth

PK,FK2 W PK N PK,FK1 P
PK,FK2 C
FK1 H S
A

Marriage Person Birth

PK,FK1 H PK N PK,FK1 P
PK,FK2 C
FK2 W S
A

 Because we want to focus on relational algebra, which

does not understand keys, we will not specify keys in
this unit

74
Microsoft Access Database

 Microsoft Access Database with this example

has been posted
» The example suggests that you download and install
Microsoft Access 2007
» The examples are in the Access 2000 format so that
if you have an older version, you can work with it
 Access is a very good tool for quickly learning
basic constructs of SQL DML, although it is not
suitable for anything other than personal
databases

75
Microsoft Access Database

 The database and our queries (other than the one with
MINUS at the end) are the appropriate “extras” directory
on the class web in “slides”
» MINUS is frequently specified in commercial databases in a
roundabout way
» We will cover how it is done when we discuss commercial
databases
 Our sample Access database: People.mdb
 The queries in Microsoft Access are copied and pasted
in these notes, after reformatting them
 Included copied and pasted screen shots of the results
of the queries so that you can correlate the queries with
the names of the resulting tables

76
Our Database With Sample Queries - Open In Microsoft Access

77
Our Database

Person N S A Birth P C
Albert M 20 Dennis Albert
Dennis M 40 John Mary
Evelyn F 20 Mary Albert
John M 60 Robert Evelyn
Mary F 40 Susan Evelyn
Robert M 60 Susan Richard
Susan F 40

Marriage H W A
Dennis Mary 20
Robert Susan 30

78
Our Instance In Microsoft Access

79
A Query

 Produce the relation Answer(A) consisting of all

ages of people
 Note that all the information required can be
obtained from looking at a single relation,
Person
 Answer:= A
SELECT A 20
40
FROM Person; 20
60
40
60
40
80
The Query In Microsoft Access

 The actual query was copied and pasted

from Microsoft Access and reformatted for
readability
 The result is below

81
A Query

 Produce the relation Answer(N) consisting of all

women who are less or equal than 32 years old.
 Note that all the information required can be
obtained from looking at a single relation,
Person
 Answer:=
SELECT N
FROM Person
WHERE A <= 32 AND S =‘F’;

N
Evelyn
82
The Query In Microsoft Access

 The actual query was copied and pasted

from Microsoft Access and reformatted for
readability
 The result is below

83
A Query

 Produce a relation Answer(P, Daughter) with

the obvious meaning
 Here, even though the answer comes only from
the single relation Birth, we still have to check in
the relation Person what the S of the C is
 To do that, we create the Cartesian product of
the two relations: Person and Birth. This gives
us “long tuples,” consisting of a tuple in Person
and a tuple in Birth
 For our purpose, the two tuples matched if N in
Person is C in Birth and the S of the N is F

84
A Query

Answer:=

SELECT P, C AS Daughter
FROM Person, Birth
WHERE C = N AND S = ‘F’;

P Daughter
John Mary
Robert Evelyn
Susan Evelyn
85
Cartesian Product With Condition: Matching Tuples Indicated

Person N S A
Albert M 20
Dennis M 40
Evelyn F 20
John M 60
Mary F 40
Robert M 60
Susan F 40

Birth P C
Dennis Albert
John Mary
Mary Albert
Robert Evelyn
Susan Evelyn
Susan Richard

86
The Query In Microsoft Access

 The actual query was copied and pasted

from Microsoft Access and reformatted for
readability
 The result is below

87
A Query

 Produce a relation Answer(Father, Daughter)

with the obvious meaning.
 Here we have to simultaneously look at two
copies of the relation Person, as we have to
determine both the S of the Parent and the S of
the C
 We need to have two distinct copies of
Person in our SQL query
 But, they have to have different names so we
can specify to which we refer
 Again, we use AS as a renaming operator,
these time for relations
88
A Query

 Answer :=
SELECT P AS Father, C AS Daughter
FROM Person, Birth, Person AS Person1
WHERE P = Person.N AND C = Person1.N
AND Person.S = ‘M’ AND Person1.S = ‘F’;

Father Daughter
John Mary
Robert Evelyn

89
Cartesian Product With Condition: Matching Tuples Indicated

Person N S A Person N S A
Albert M 20 Albert M 20
Dennis M 40 Dennis M 40
Evelyn F 20 Evelyn F 20
John M 60 John M 60
Mary F 40 Mary F 40
Robert M 60 Robert M 60
Susan F 40 Susan F 40

Birth P C
Dennis Albert
John Mary
Mary Albert
Robert Evelyn
Susan Evelyn
Susan Richard

90
The Query In Microsoft Access

 The actual query was copied and pasted

from Microsoft Access and reformatted for
readability
 The result is below

91
A Query

 Produce a relation: Answer(Father_in_law,

Son_in_law).
 A classroom exercise, but you can see the
solution in the posted database.
 Hint: you need to compute the Cartesian
product of several relations if you start from
scratch, or of two relations if you use the
previously computed (Father, Daughter) relation

F_I_L S_I_L
John Dennis

92
The Query In Microsoft Access

 The actual query was copied and pasted

from Microsoft Access and reformatted for
readability
 The result is below

93
A Query

 Produce a relation:
Answer(Grandparent,Grandchild)
 A classroom exercise, but you can see
the solution in the posted database

G_P G_C
John Albert

94
Cartesian Product With Condition: Matching Tuples Indicated

Birth P C Birth P C
Dennis Albert Dennis Albert
John Mary John Mary
Mary Albert Mary Albert
Robert Evelyn Robert Evelyn
Susan Evelyn Susan Evelyn
Susan Richard Susan Richard

95
The Query In Microsoft Access

 The actual query was copied and pasted

from Microsoft Access and reformatted for
readability
 The result is below

96
Further Distance

 How to compute (Great-grandparent,Great-grandchild)?

 Easy, just take the Cartesian product of the
(Grandparent, Grandchild) table with (Parent,Child)
table and specify equality on the “intermediate” person
 How to compute (Great-great-grandparent,Great-great-
grandchild)?
 Easy, just take the Cartesian product of the
(Grandparent, Grandchild) table with itself and specify
equality on the “intermediate” person
 Similarly, can compute (Greatx-grandparent,Greatx-
grandchild), for any x
 Ultimately, may want (Ancestor,Descendant)

97
Relational Algebra Is Not Universal:Cannot Compute (Ancestor,Descendant)

 Standard programming languages are

universal
 This roughly means that they are as powerful as
Turing machines, if unbounded amount of
storage is permitted (you will never run out of
memory)
 This roughly means that they can compute
anything that can be computed by any
computational machine we can (at least
currently) imagine
 Relational algebra is weaker than a standard
programming language
 It is impossible in relational algebra (or standard
SQL) to compute the relation Answer(Ancestor,
Descendant)
98
Relational Algebra Is Not Universal: Cannot Compute (Ancestor,Descendant)

 It is impossible in relational algebra (or standard SQL)

to compute the relation Answer(Ancestor, Descendant)
 Why?
 The proof is a reasonably simple, but uses cumbersome
induction.
 The general idea is:
» Any relational algebra query is limited in how many relations or
copies of relations it can refer to
» Computing arbitrary (ancestor, descendant) pairs cannot be
done, if the query is limited in advance as to the number of
relations and copies of relations (including intermediate results)
it can specify
 This is not a contrived example because it shows that
we cannot compute the transitive closure of a directed
graph: the set of all paths in the graph

99
A Sample Query

 Produce a relation Answer(A) consisting

of all ages of visitors that are not ages of
marriages

SELECT
A FROM Person
MINUS
SELECT
A FROM MARRIAGE;

100
The Query In Microsoft Access

 We do not show this here, as it is done in

a roundabout way and we will do it later

101
It Does Not Matter If We Remove Duplicates

 Removing duplicates
A A
A
20
 40
- 20
30
= 40
60
60

 Not removing duplicates

A
20 A
A
40 40
20
20
- 30 = 60
60 40
40 60
60 40
40
102
It Does Not Matter If We Remove Duplicates

 The resulting set contains precisely ages: 40, 60

 So we do not have to be concerned with whether the
implementation removes duplicates from the result or not
 In both cases we can answer correctly
» Is 50 a number that is an age of a marriage but not of a person
» Is 40 a number that is an age of a marriage but not of a person
 Just like we do not have to be concerned with whether it
sorts (orders) the result
 This is the consequence of us not insisting that an
element in a set appears only once, as we discussed
earlier
 Note, if we had said that an element in a set appears
once, we would have to spend effort removing
duplicates!
103
Now To “Pure” Relational Algebra

 This was described in several slides

 But it is really the same as before, just the
notation is more mathematical
 Looks like mathematical expressions, not
snippets of programs
 It is useful to know this because many
articles use this instead of SQL
 This notation came first, before SQL was
invented, and when relational databases
were just a theoretical construct
104
p: Projection: Choice Of Columns

R A B C D
1 10 100 1000
1 20 100 1000
1 20 200 1000

 SQL statement Relational

Algebra B A D
SELECT B, A, D 10 1 1000 p
B,A,D (R)
FROM R 20 1 1000
20 1 1000

 We could have removed the duplicate row,

but did not have to 105
s: Selection: Choice Of Rows
R A B C D
5 5 7 4
5 6 5 7
4 5 4 4
5 5 5 5
4 6 5 3
4 4 3 4
4 4 4 5
4 6 4 6
 SQL statement: Relational Algebra
SELECT * sA  C  D=4 (R) Note: no need for p
FROM R
WHERE A <= C AND D = 4;
A B C D
5 5 7 4
4 5 4 4
106
Selection

 In general, the condition (predicate) can be

specified by a Boolean formula with
  and  on atomic conditions, where a
condition is:
» a comparison between two column names,
» a comparison between a column name and a
constant
» Technically, a constant should be put in quotes
» Even a number, such as 4, perhaps should be put in
quotes, as ‘4’ so that it is distinguished from a
column name, but as we will never use numbers for
column names, this not necessary

107
: Cartesian Product
R A B S C B D
1 10 40 10 10
2 10 50 20 10
2 20
 SQL statement Relational Algebra
SELECT A, R.B, C, S.B, D RS
FROM R, S
A R.B C S.B D
1 10 40 10 10
1 10 50 20 10
2 10 40 10 10
2 10 50 20 10
2 20 40 10 10
2 20 50 20 10
108
A Typical Use Of Cartesian Product

R Size Room# S ID# Room# YOB

140 1010
150 1020 40 1010 1982
140 1030 50 1020 1985

 SQL statement: Relational Algebra

SELECT ID#, R.Room#, Size p ID#, R.Room#, Size (sR.B = S.B (R  S))
FROM R, S
WHERE R.Room# = S.Room#

ID# R.Room# Size

40 1010 140
50 1020 150
109
: Union

R A B S A B
1 10 1 10
2 20 3 20

 SQL statement Relational Algebra

(SELECT * RS
A B
FROM R)
1 10
UNION
(SELECT * 2 20
FROM S) 3 20
 Note: We happened to choose to remove duplicate
rows
 Union compatibility required
110
: Difference

R A B S A B
1 10 1 10
2 20 3 20

 SQL statement Relational

Algebra
(SELECT * A B RS
FROM R) 2 20
MINUS
(SELECT *
FROM S)
 Union compatibility required
111
: Intersection

R A B S A B
1 10 1 10
2 20 3 20

 SQL statement Relational

Algebra A B
(SELECT * RS
FROM R) 1 10
INTERSECT
(SELECT *
FROM S)
 Union compatibility required
112
Summary (1/2)

 A relation is a set of rows in a table with labeled columns

 Relational algebra as the basis for SQL
 Basic operations:
» Union (requires union compatibility)
» Difference (requires union compatibility)
» Intersection (requires union compatibility); technically not a basic
operation
» Selection of rows
» Selection of columns
» Cartesian product
 These operations define an algebra: given an expression
on relations, the result is a relation (this is a “closed”
system)
 Combining this operations allows production of
sophisticated queries
113
Summary (2/2)

 Relational algebra is not universal: We can write

programs producing queries that cannot be
specified using relational algebra
 We focused on relational algebra specified
using SQL syntax, as this is more important in
practice
 The other, “more mathematical” notation came
first and is used in research and other venues,
but not commercially

114
Agenda

1 Session Overview

2 Relational Algebra and Relational Calculus

3 Relational Algebra Using SQL Syntax

4 Summary and Conclusion

115
Summary

 Relational algebra and relational calculus are formal

languages for the relational model of data
 A relation is a set of rows in a table with labeled columns
 Relational algebra and associated operations are the
basis for SQL
 These relational operations define an algebra
 Relational algebra is not universal as it is possible to write
programs producing queries that cannot be specified
using relational algebra
 Relational algebra can be specified using SQL syntax
 The other, “more mathematical” notation came first and is
used in research and other venues, but not commercially

116
Assignments & Readings

 Readings
» Slides and Handouts posted on the course web site
» Textbook: Chapters 6

 Assignment #3
» Textbook exercises: Textbook exercises: 9.3, 9.5, 10.23, 6.16, 6.24, 6.32

 Project Framework Setup (ongoing)

117
Next Session: Standard Query Language (SQL) Features

 SQL as implemented in commercial

databases

118
Any Questions?

119