Relations and Relational Algebra

CSCI1270, Fall 2009, Lecture 1

Introduced by Ted Codd (late 60’s – early 70’s)
● Before = “Network Data Model” (Cobol as DDL,
DML)
● Very contentious: Database Wars
Relational data model
contributes:
1.Separation of logical and physical data models (data
independence)
2.Declarative query languages
3.Formal semantics
First 4.Query optimization (key to commercial success)
● Ingres (UC Berkeley)
prototypes:
● System R (IBM)•
DB2
CSCI1270, Fall 2008, Lecture 2
 bname acct_no balance
Downtown A-101 500
Account =
Brighton A-201 900
Brighton A-217 500
Table name

Attribute
names
Terms:
● Tables • Relations
● Columns • Attributes
● Rows • Tuples
● Schema (e.g.: Acct_Schema = (bname, acct_no,
balance))
CSCI1270, Fall 2008, Lecture 2
Mathematical
relations
Given sets: R = {1, 2, 3}, S = {3, 4}
● R  S = { (1, 3), (1, 4), (2, 3), (2, 4), (3, 3), (3, 4) }
● A relation on R, S is any subset () of R  S (e.g: { (1, 4), (3, 4)})

Database
Given attribute domains
relations
Branches = { Downtown, Brighton, … }
Accounts = { A-101, A-201, A-217, … }
Balances = R – real numbers
Account  Branches  Accounts  Balances
{ (Downtown, A-101, 500),
(Brighton, A-201, 900),
(Brighton, A-217, 500) }
CSCI1270, Fall 2008, Lecture 2
 bname acct_no balance
Downtown A-101 500
Account =
Brighton A-201 900
Brighton A-217 500

Considered equivalent to…

{ (Downtown, A-101, 500),
(Brighton, A-201, 900),
(Brighton, A-217, 500) }

Relational database semantics are defined in

terms of mathematical relations

CSCI1270, Fall 2008, Lecture 2

 Account Branch
bname acct_no balance bname bcity assets

Depositor
cname acct_no Borrower
cname lno

Customer
cname cstreet ccity
Loan
bname lno amt

CSCI1270, Fall 2008, Lecture 2

 Account Branch
bname acct_no balance bname bcity assets
Downtown A-101 500 Downtown Brooklyn 9M
Mianus A-215 700 Redwood Palo Alto 2.1M
Perry A-102 400 Perry Horseneck 1.7M
R.H. A-305 350 Mianus Horseneck 0.4M
Brighton A-201 900 R.H. Horseneck 8M
Redwood A-222 700 Pownel Bennington 0.3M
Brighton A-217 750 N. Town Rye 3.7M
Brighton Brooklyn 7.1M
Depositor
cname acct_no Borrower
Johnson A-101 cname lno
Smith A-215
Hayes A-102 Jones L-17
Turner A-305 Smith L-23
Johnson A-201 Hayes L-15
Jones A-217 Jackson L-14
Lindsay A-222 Curry L-93
Smith L-11
Customer Williams L-17
Adams L-16
cname cstreet ccity
Loan
Jones Main Harrison
Smith North Rye bname lno amt
Hayes Main Harrison
Downtown L-17 1000
Curry North Rye
Redwood L-23 2000
Lindsay Park Pittsfield
Perry L-15 1500
Turner Putnam Stanford
Downtown L-14 1500
Williams Nassau Princeton
Mianus L-93 500
Adams Spring Pittsfield
R.H. L-11 900
Johnson Alma Palo Alto
Perry L-16 1300
Glenn Sand Hill Woodside
Brooks Senator Brooklyn
Green Walnut Stanford

CSCI1270, Fall 2008, Lecture 2

Kinds of
● 1. Superkeys
keys ○ set of attributes of table for which every row has distinct set of values
○ in example below: (bname,bcity) is a superkey
● 2. Candidate keys
○ “minimal” superkeys
● 3. Primary keys
○ DBA-chosen candidate keys

Act as Integrity
i.e., guard against illegal/invalid instance of given
Constraints
schema
e.g., Branch
bname = (bname, assets 
bcity bcity, assets)
Brighton Brooklyn 5M
Brighton Boston 3M
Invalid!!

CSCI1270, Fall 2008, Lecture 2

Recall: Query = “Retrieval Program”
Language
Examples:
Theoretical:
1.Relational Algebra
2.Relational Calculus
1.Tuple Relational Calculus (TRC)
2.Domain Relational Calculus (DRC)
Practical:
1.SQL (originally: SEQUEL from System R)
2.Quel (used in Ingres)
3.Datalog (Prolog-like – used in research lab systems)

● Theoretical QLs give semantics to Practical QLs

CSCI1270, Fall 2008, Lecture 2

●Basic Operators
1.select ( σ )
2.project (  )
3.union (  )
4.set difference ( – )
5.cartesian product (  )
6.rename ( ρ )

●Closure Property
Relatio
n Relational Relatio
n Relational
Operator Operator
Relatio
n

CSCI1270, Fall 2008, Lecture 2

Notation: σpredicate (Relation)
Relation: Can be name of table, or the result of another query
Predicate:
1.Simple
○ attribute1 = attribute2
○ attribute = constant value (also: ≠, <, >, ≤, ≥)

2. Complex
○ predicate AND predicate
○ predicate OR predicate
○ NOT predicate

Idea
: Select rows from a relation based on a
predicate

CSCI1270, Fall 2008, Lecture 2

Notation: σpredicate (Relation)

bname bcity assets

σ bcity = “Brooklyn” (branch) = Downtown Brooklyn 9M
Brighton Brooklyn 7.1M

σ assets > $8M (σ bcity = “Brooklyn” (branch)) =

bname bcity assets
Downtown Brooklyn 9M

CSCI1270, Fall 2008, Lecture 2

Notation: A1, …, An (Relation)
● Relation: name of a table or result of another
query
● Each Ai is an attribute
● Idea:  selects columns (vs. σ which selects rows)
cstreet ccity
 cstreet, ccity (customer) = Main Harrison
North Rye
Park Pittsfield
Putnam Stanford
Nassau Princeton
Spring Pittsfield
Alma Palo Alto
Sand Hill Woodside
Senator Brooklyn
Walnut Stanford

CSCI1270, Fall 2008, Lecture 2

Notation: A1, …, An (Relation)
● Each Ai an attribute
● Idea:  selects columns (vs. σ which selects
rows)

bcity
bcity (σassets > 5M (branch))
Brooklyn
= Horseneck

CSCI1270, Fall 2008, Lecture 2

Notation: Relation1  Relation2
R  S valid only if:

1.R, S have same number of columns (arity)

2.R, S corresponding columns have same domain (compatibility)

Example:
cname
Johnson
(cname (depositor))  (cname (borrower)) = Smith
Depositor Borrower Hayes
cname acct_no cname lno Turner
Jones
Lindsay
Jackson
Curry
Williams
Adams

CSCI1270, Fall 2008, Lecture 2

Notation: Relation1 - Relation2
R - S valid only if:

1.R, S have same number of columns (arity)

2.R, S corresponding columns have same domain (compatibility)

Example:
( bname (σamount ≥ 1000 (loan))) – (bname (σ balance < 800 (account)))
Loan Account

bname lno amt = bname acct_no balance

Downtown A-101 500
Downtown L-17 1000
Mianus A-215 700
Redwood L-23 2000
Perry L-15 1500 bname acct_no balance Perry A-102 400
R.H. A-305 350
Downtown L-14 1500
Mianus L-93 500 Mianus A-215 700 Brighton A-201 900
Redwood A-222 700
R.H. L-11 900 Brighton A-201 900 Brighton A-217 750
Perry L-16 1300
Redwood A-222 700
Brighton A-217 750

CSCI1270, Fall 2008, Lecture 2

Notation: Relation1 - Relation2
R - S valid only if:

1.R, S have same number of columns (arity)

2.R, S corresponding columns have same domain (compatibility)

Example:
( bname (σamount ≥ 1000 (loan))) – (bname (σ balance < 800 (account)))
bname lno amount
=
bname acct_no balance
bname
Downtown L-17 1000 Mianus A-215 700
Downtown
Redwood L-23 –
2000 Brighton A-201 900
Perry L-15 1500 Redwood
=
A-222 700
Perry

Downtown L-14 1500 Brighton A-217 750

Perry L-16 1300

CSCI1270, Fall 2008, Lecture 2

Notation: Relation1  Relation2
R  S like cross product for mathematical relations:
● every tuple of R appended to every tuple of
S
Example:
depositor. acct_no borrower. lno
depositor  borrower = cname cname
Johnson A-101 Jones L-17
Johnson A-101 Smith L-23
How many tuples in Johnson A-101 Hayes L-15
Johnson A-101 Jackson L-14
the result? Johnson A-101 Curry L-93
Johnson A-101 Smith L-11
Johnson A-101 Williams L-17
A: 56 Johnson A-101 Adams L-16
Smith A-215 Jones L-17
… … … …

CSCI1270, Fall 2008, Lecture 2

Notation:  identifier (Relation)
renames a relation, or

Notation:  identifier0 (identifier1, …, identifiern) (Relation)

renames relation and columns of n-column relation

Use:
massage relations to make , – valid, or  more readable

CSCI1270, Fall 2008, Lecture 2

Notation:  identifier0 (identifier1, …, identifiern) (Relation)
renames relation and columns of n-column relation
Example:
 res (dcname, acctno, bcname, lno) (depositor  borrower) =
depositor. acct_no borrower. lno
cname cname
Johnson A-101 Jones L-17
Johnson A-101 Smith L-23
Johnson A-101 Hayes L-15
Johnson A-101 Jackson L-14
Johnson A-101 Curry L-93
Johnson A-101 Smith L-11
Johnson A-101 Williams L-17
Johnson A-101 Adams L-16
Smith A-215 Jones L-17
… … … …

CSCI1270, Fall 2008, Lecture 2

●Determine lno for loans that are for an amount that
is larger than the amount of some other loan. (i.e.
lno for all non-minimal loans)
Can do in
steps:
Temp1 • …
Temp2 • … Temp1 …
…

CSCI1270, Fall 2008, Lecture 2

1. Find the base data we need
lno amt
L-17 1000
Temp1 • lno,amt (loan) L-23 2000
L-15 1500
L-14 1500
L-93 500
L-11 900
L-16 1300

2. Make a copy of (1)

lno2 amt2
Temp2 • ρ Temp2 (lno2,amt2) (Temp1) L-17 1000
L-23 2000
L-15 1500
L-14 1500
L-93 500
L-11 900
L-16 1300

CSCI1270, Fall 2008, Lecture 2

3. Take the cartesian product of 1 and 2

Temp3 • Temp1  Temp2

lno amt lno2 amt2
L-17 1000 L-17 1000
L-17 1000 L-23 2000
… … … …
L-17 1000 L-16 1300
L-23 2000 L-17 1000
L-23 2000 L-23 2000
… … … …
L-23 2000 L-16 1300
… … … …

CSCI1270, Fall 2008, Lecture 2

4. Select non-minimal loans

Temp4 • σamt > amt2 (Temp3)

5. Project on lno

Result •  lno (Temp4)

… or, if you prefer…

● lno (σamt > amt2 (
○ lno,amt (loan)  (
○ ρTemp2 (lno2,amt2) (lno,amt (loan)))))

CSCI1270, Fall 2008, Lecture 2

Relational Algebra
1.SELECT ( σ )
2.PROJECT ( π )
3.UNION (  )
4.SET DIFFERENCE ( – )
5.CARTESIAN PRODUCT (  )
6.RENAME ( ρ )

● Relational algebra gives semantics to practical query languages

● Above set: minimal relational algebra
• will now look at some redundant (but useful!) operators

CSCI1270, Fall 2008, Lecture 2

Express the following query in the
RA:
Find the names of customers who have both accounts and
loans
T1 • ρT1 (cname2, lno) (borrower)

T2 • depositor  T1
T3 • σcname = cname2 (T2)
Result
Above sequence • π cname (ρ,
of operators (T3,
) σ) very
common.
Motivates additional (redundant) RA operators.

CSCI1270, Fall 2008, Lecture 2

 1.Natural Join ( )
2. Division ( )
3. Generalized Projection (π)
4. Outer Joins ( )
5. Update ( • ) (we’ve already been using)
● Redundant: Above can be expressed in terms of minimal RA
• e.g. depositor borrower =
π …(σ…(depositor  ρ…(borrower)))
● Added for convenience
CSCI1270, Fall 2008, Lecture 2
Notation: Relation Relation
1 2
Idea: combines ρ, ,
σ
A B C D A B C D E
E B D
1 α + 10 ‘a’ α 10 1 α + 10 ‘a’
2 α - 10 ‘a’ α 20 = 2 α - 10 ‘a’
2 α - 20 ‘b’ β 10 2 α - 20 ‘a’
3 β + 10 ‘c’ β 10 3 β + 10 ‘b’

r s

deposito borrower
r
≡
πcname,acct_no,lno (σcname=cname2 (depositor  ρt(cname2,lno) (borrower)))

CSCI1270, Fall 2008, Lecture 2

Notation: Relation Relation
1 2
Idea: expresses “for all” queries
A B
α 1
α 23 B A
α 1 1 = αδ
β 1 2
γ 3
γ 4 s
γ 6
γ 1
δ 2
Query: Find values for A in r
δ which have corresponding B
r values for all B values in s
CSCI1270, Fall 2008, Lecture 2
 
Another way to look at it: and 
17 3 = The largest value of i such

5 that: i  3 ≤ 17
Relational
A B
Division α 1
α 23 B A
α 1 1 α
=
β 1 2 δ
γ 3
s t
γ 4
γ 6
γ 1 The largest value of t such
δ 2 that:
δ
r (tsr)
CSCI1270, Fall 2008, Lecture 2
A More Complex Example

A B C D E
α a α a 1
α a γ a 1 A B C
D E
α a γ b 1 = α ?a γ
β a γ a 1 a 1
γ a γ
β a γ b 3 b 1
γ a γ a 1 s t
γ a γ b 1
γ a β b 1
r

CSCI1270, Fall 2008, Lecture 2

Notation: e1,…,en
(Relation)
e1,…,en can include arithmetic expressions – not just
attributes
Exampl
e cname limit balance

credit = Jones 5000 2000

Turner 3000 2500

Then… cname limit-balance

Jones 3000
π cname, limit - balance (credit) =
Turner 500

CSCI1270, Fall 2008, Lecture 2

 ● Aggregation function takes a collection of values and
returns a single value as a result.
● avg: average value
min: minimum value
max: maximum value
sum: sum of values
count: number of values
● Aggregate operation in relational algebra

G1, G2, …, Gn g F1( A1), F2( A2),…, Fn( An) (E)

● E is any relational-algebra expression
● G1, G2 …, Gn is a list of attributes on which to group (can be empty)
● Each Fi is an aggregate function
● Each Ai is an attribute name

CSCI1270, Fall 2008, Lecture 2

● Relation r:
A B C

  7
  7
  3
  10

sum-C
g sum(c) (r) 27

CSCI1270, Fall 2008, Lecture 2

● Relation account grouped by branch-name:
branch-name account- balance
Perryridge number
A-102 400
Perryridge A-201 900
Brighton A-217 750
Brighton A-215 750
Redwood A-222 700

branch-name g sum(balance) (account)

branch-name balance
Perryridge 1300
Brighton 1500
Redwood 700
CSCI1270, Fall 2008, Lecture 2
●Result of aggregation does not have a name
○ Can use rename operation to give it a name
○ For convenience, we permit renaming as part of
aggregate operation
branch-name g sum(balance) as sum-balance (account)

CSCI1270, Fall 2008, Lecture 2

Motivation:
bname lno amt cname lno
Downtown L-170 3000 Jones L-170
loan = borrower =
Redwood L-230 4000 Smith L-230
Perry L-260 1700 Hayes L-155

=
bname lno amt cname
loan borrower = Downtown L-170 3000 Jones
Redwood L-230 4000 Smith

Join result loses…

• any record of Perry
• any record of Hayes
CSCI1270, Fall 2008, Lecture 2
 bname lno amt cname lno

loan = Downtown L-170 3000 borrower = Jones L-170

Redwood L-230 4000 Smith L-230
Perry L-260 1700 Hayes L-155

1. Left Outer Join ( )

● preserves all tuples in left
relation
loan borrower =

bname lno amt cname

Downtown L-170 3000 Jones
Redwood L-230 4000 Smith ┴ = NULL
Perry L-260 1700 ┴

CSCI1270, Fall 2008, Lecture 2

 bname lno amt cname lno

loan = Downtown L-170 3000 borrower = Jones L-170

Redwood L-230 4000 Smith L-230
Perry L-260 1700 Hayes L-155

2. Right Outer Join ( )

● preserves all tuples in right
relation
loan borrower =

bname lno amt cname

Downtown L-170 3000 Jones
Redwood L-230 4000 Smith ┴ = NULL
┴ L-155 ┴ Hayes

CSCI1270, Fall 2008, Lecture 2

 bname lno amt cname lno

loan = Downtown L-170 3000 borrower = Jones L-170

Redwood L-230 4000 Smith L-230
Perry L-260 1700 Hayes L-155

3. Full Outer Join ( )

● preserves all tuples in both
relations
loan borrower =
bname lno amt cname
Downtown L-170 3000 Jones
Redwood L-230 4000 Smith
Perry L-260 1700 ┴
┴ L-155 ┴ Hayes ┴ = NULL

CSCI1270, Fall 2008, Lecture 2

Notation: Identifier • Query
Common Uses:
1. Deletion: r • r – s
e.g., account • account – σbname=Perry (account)
(deletes all Perry
accounts)r • r  s
2. Insertion:
e.g., branch • branch  {(Waltham, Boston, 7M)}
(inserts new branch with
bname = Waltham, bcity = Boston, assets = 7M)
e.g., depositor • depositor  (ρtemp (cname,acct_no) (borrower))
(adds all borrowers to depositors, treating lno’s as acct_no’s)
3. Update: r • πe1,…,en (r)
e.g., account • πbname,acct_no,bal*1.05 (account)
(adds 5% interest to account balances)
CSCI1270, Fall 2008, Lecture 2