Functional Dependencies

FUNCTIONAL DEPENDENCIES
CSC343
Functional Dependencies Functional Dependencies
Examples: PERSON Relation
PERSON SIN NAME CITY

123 Laurent Toronto
324 Bill Toronto
574 Bill Montreal
What can we say about Person table?
“If I know the sin number I know the name”

SIN attribute determines NAME attribute.
Attribute NAME functionally depends on attribute SIN
Warning: Knowing the NAME does not imply the SIN knowledge: NAME 6! SIM
NOTATION: SIN ! NAME
CSC343
COURSES Relation
STUDENT Name
0,n
enroll
0,n
COURSES Name
1,1
teach
0,n
PROF Name
COURSE NAME PROF STUDENT

Database MIGNET SMITH
Database MIGNET BILL
Database MIGNET SMITH
Math HARDIN GEORGE
“A Course has only one Professor”

NAME ! PROF
2 tuples that have the same value for NAME have the same value for PROF
3
CSC343
Key Example
PERSON (SIN, LastName, FirstName, Address)
SIN ! LastName
SIN ! FirstName
SIN ! Address
If we know the SIN value, we know all the attributes.

2 tuples sharing the same SIN are identical
SIN identifies a tuple.
SIN is a key
SIN LastName ! FirstName
SIN LastName is a superkey (LastName is redundant)
CSC343
SUPPLYING Relation
SUPP (SNAME, PNAME, QUANT, PRICE)
SNAME PNAME ! QUANT

SNAME PNAME ! PRICE
SNAME PNAME is a key.
SUPPLIER Sname
0,n
Quant SUPPLYING Price
0,n
PRODUCT Pname
CSC343
ADDRESS Relation
ADDRESS (STREET, NUMBER, CITY, ZIPCODE)
Simplifying Hypothesis
Several zipcode for a city: Toronto = M5B3F4, M5S2E4, . . .

One city for a zipcode: Mxxxxx = Toronto
An address (street + number + city) belongs only to one zipcode: 4 St George Street,
Toronto = M5S 2E4
Number Street City ! Zipcode

Zipcode ! City
Keys of the relation ADDRESS: (Street Number City) and (Number Street
Zipcode)
CSC343
Definitions
a) Functional Dependencies
Let R(U) a relational schema, r one relation of the schema R, X U, Y U two
attribute sub-set of R. The functional dependencies
X ! Y
is true in r, iff (if and only if) for every tuples of r that share the same value for all
attributes of X, they also share the same value for all attributes of Y.
Example:
ADDRESS(STREET, NUMBER, CITY, ZIPCODE)
X = ZipCode
Y = City
CSC343
b) SuperKey
Let R(U) a schema and X U a attribute subset.
X is a superkey superkey of r of the schema R, if X ! U.
Example:
SIN LASTNAME ! SIN LASTNAME FIRSTNAME ADDRESS
SIN LASTNAME is a superkey
c) KEY
X is a key, if:
1. X is a superkey: X ! U
2. it does not exist Y X, such that Y ! U
Example:
STREET NUMBER CITY ! STREET NUMBER CITY ZIPCODE
STREET NUMBER CITY is a key (why?)
CSC343
Finding a Key: Example
COURSE(Name,Hour,Room,Prof)
F =N ! HR, HR ! P
We can proof from F that if we know the name of the course, we also know the name of
the professor. (Name is a key):
1. N ! HR: two tuples that share the name of the course share also the hour and the
room value.
2. HR ! P: two tuples that share the hour and the room share also the name of the
professor.
3. FD 1 and 2 implie that two tuples which have the same name for the course have
also the same professor name: N ! P
We can define some properties and rules on the FD which permit to deduct others FDs.
CSC343
Functional Dependencies Properties
(Armstrong’s Axioms)
1) Reflexivity
If X Y then Y ! X (for every attribute subsets X et Y)
Example:
NAME CITY ! NAME
Trivial: “Two persons who have the same name and live in the same city have the
same name.”
10
CSC343
2) Transitivity
If X ! Y and Y ! Z, then X ! Z
Example :
R(SIN, ZIPCODE, CITY)
fSIN ! ZIPCODE, ZIPCODE ! CITYg j= SIN ! CITY
“If we know the zipcode from a sin number and the city form the zipcode then we
know the city from the sin number.”
3) Augmentation
X ! Y j= XZ ! YZ
Example:
SIN ! ZIPCODE j= SIN CITY ! ZIPCODE CITY
11
CSC343
4) Union et decomposition
fX ! A, X ! Bg,X ! AB
5) Pseudotransitivity
fX ! Y, WY ! Z g j= WX ! Z
REMARK: Union, decomposition and pseudotransitivity can be deduced from the others
axioms.
12
CSC343
Closing set of a functional dependencies set
From the functional dependencies set given from the real world and using the preceding
properties (Armstrong’s axioms) we can deduce others FDs:
Examples:
1) R(A,B,C,D)
F = f A ! B, B ! C g
Transitivity: A ! C
Notation: F j= A ! C
13
CSC343
2) R(Course, Prof, Hour, Room, Student, Mark)

F = f C ! P, HR ! C, HP ! R, CS ! M, HS ! R g
F j= HR ! P, HS ! C,P,M
So HS is a key Why?
The union of F and every deduced FD is called closing set, or attribute closure of F is
denoted F +.
14
CSC343
Minimal Closure
The minimal closure G of a set F of functional dependencies (FD) is a set of FD such that:
1. We can infer from G the same FD than from F : G + = F +

2. Only one attribute is on the right side on every FD in G (decomposition)
3. Every FD are useful: if we delete one we can not obtain F + anymore.
4. Every FD are elemantary: the set fA !C; AC!Bg is redundant: we replace AC!B,
which is not elementary by A!B.
15
CSC343
Example
Functional Minimal Closure

Dependencies
AB ! C AB ! C
C!A C!A
BC ! D BC ! D
ACD ! B CD ! B
D ! EG D!E
BE ! C D!G
CE ! AG BE ! C
CE ! G
CE ! A is missing, why?
ACD ! B is not elementary, why?
16
CSC343
Answer:
ACD ! B is not elemenaty, why?
C ! A j= CCD ! ACD by augmentation
CCD ! ACD j= CD! ACD (A ! A by reflexivity and union)

i
CD ! ACD and ACD ! B j= CD ! B by transitivity
Second demonstration
By pseudotransitivity, f C ! A, ACD ! Bg j= CCD ! B j= CD ! B.
17
CSC343
Minimal Closure
Concretely, the minimal closure
is either directly given by the real world analysis.

is eitheir trivialy deduced because:
– every FD are useful and elementary
– It is enough to decompose the right side of every FDs.
18
CSC343
New Definition for a SuperKey
Let R(U) a relational schema, F a set of functional dependencies and X

U a set of attributes. X is a superkey of R if for all A 2 U
F j= X ! A
or if:
X ! A 2 F+
Compute F + from F might take time.
But show that F j= X ! A is easy and fast.
19
CSC343
Calculation of a key of a relation
Let R(U) a relational schema and X U a set of attributes.
We define X + as a subset of attributes A such that F j= X ! A

If A belongs to X +, then by definition F j= X ! A
X + is the set of functionally dependent attributes of X.
To find a key we use the following algorithm:
1. We search for one X such that X + = U ) X is a superkey

2. X is a key, if it does not exist Y X such that Y + = U
20
CSC343
X+
Step 1: We start from X

For every Y ! A, such that Y X, Y ! A 2 F , we add A to X.
We get X 1.
Step i: We start X 1 i
For every Y ! A, such that Y X 1, Y

i
! A 2 F , we add A to X i 1.
We get X . i
We stop when we do not find new FDs anymore.
X + = X +1 = X
i i
21
CSC343
Examples
1. Show that HS is a key for R(CHSNRP) with the set F of FD given before.
2. for the relation ADDRESS(CITY, STREET, NUMBER, ZIPCODE), show
that CITY STREET NUMBER is a key. What is the other one?
22
CSC343
UPDATE ANOMALIES
CSC343
Functional Dependencies Update Anomalies
Example
Let the schema S1:
Supplier(SNAME, FADDRESS)
Product(SNAME, PNAME, PRICE)
and the set of DF: F = fSNAME ! FADDRESS, (SNAME PNAME) ! PRICEg
Suppose now that we replace S1 by S2:
R(SNAME, SADDRESS, PNAME, PRICE)
24
CSC343
Anomalies
R(SNAME, SADDRESS, PNAME, PRICE)
F = fSNAME ! SADDRESS, (SNAME PNAME) ! PRICEg
What is the key of R?
SMITH Toronto COMPUTER 1000

JONS Montreal COMPUTER 900
SMITH Toronto KEYBOARD 400
1) REDONDANCY: the address for one supplier appears several times.

2) UPDATE: if we change the address in one tuple, we must also perform the same update in
the others.
25
CSC343
3) DELETION: if JONS do not supply COMPUTER anymore, we delete the second tuple we
lost any information about JONS.
4) INSERTION: we can not insert a new supplier and its address if we do not know, at least,
one product that it supplies.
SMITH Toronto COMPUTER 1000
JONS Montreal COMPUTER 900
SMITH Toronto KEYBOARD 400
DURAND NICE
=) THE INITIAL SCHEMA S1 IS “BETTER”
26
CSC343
Integrity Constraints
The list of attributes is not sufficient to describe the semantic of the real world.
It exists several types of constraints on the tuples:
1. dependencies (functional, multivalued, ...)

2. constraints that depend of the attribute domain: year < 2000
3. etc.
It is the dependencies which permits a good schema conception, i.e., the decomposition in
“good” relations.
27
CSC343
Qualities for a good schema
1. Avoid anomalies =) decomposition

2. The decomposition should keep the same amount of information
Join a relation
f1 of the schema SUPPLIER(SNAME, SADDRESS)
and a relation
f2 of the schema SUPPLIER(SNAME, PNAME, PRICE)
got by decomposition of the relation
r of the schema R(SNAME, SADDRESS, PNAME, PRICE)
must give back r.
3. The decomposition must keep the same constraints (FD). The decomposition of R in
R1(SNAME, SADDRESS, PRICE) and R2(PNAME, PRICE) do not preserve the
FD. Why?
28
CSC343
DECOMPOSITION AND NORMAL FORM
CSC343
Functional Dependencies Decomposition and Normal Form
Relation in first normal form (1NF)
All attributes are atomic (elementary)

Relation that we know.
30
CSC343
Relation 1NF
MARKS ( COURSE STUDENT MARK )

DB John 80
DB Mark 90
DB Tom 0
ARCHI Tom 100
ARCHI John 0
Relation N1FN
MARKS ( COURS PERF (STUDENT MARK) )

DB John 80
Mark 90
Tom 0
ARCHI Tom 100
John 0
31
CSC343
Relation in 3rd Normal Form
2nd Normal Form:

Purely historical.
3rd normal form: 3NF
– avoid most of the anomalies
Goal of the game: to decompose a relation (1NF) to a set of 3NF relations.
32
CSC343
3NF: First Definition
Let (R, F) be a relational schema.
We suppose that F is a minimal closure.
Definition: R is in 3NF, if for every X ! A of F ,

either X is a key
or A belongs to one of the keys.
33
CSC343
Examples: 3rd Normal Form
1) Post(City, Street, Zipcode)

F = f CS ! Z, Z ! Cg
Keys: CS, SZ
Post is in 3NF.
2) Supplier(SNAME, ADDR, PNAME, PRICE)
F = f SNAME ! ADDR, SNAME PNAME ! PRICE g
Keys: (SNAME PNAME)
Supplier is not in 3NF.
34
CSC343
3) Schedule(Course, Hour, Room)

F = f RH ! C, C ! Rg
Keys: RH, CH
Schedule is in 3NF.
4) R(A, B, C, D)
F = f AB ! C, B ! D, BC ! Ag
Keys: AB, BC
R is not in 3NF.
35
CSC343
3NF: Second Definition
Remark: It is not necessary for F to be a minimal closure. It is enough that for all FD X ! A
of F +,
A is only composed by one attribute.
A is not one of the attribute of X
The definition of a 3NF schema becames:
Definition: A relation R is in 3NF, if every FD X ! A of F + satisfies the preceding
conditions,
either X is a superkey
or A belongs to one of the keys.
Actually, it is not necessary to check all FD of F +.
It is enough to check the ones belonging to F !
36
CSC343
RemarK (Cont’d):
Why superkey in the first condition?
With conditions 1 and 2 weaker than the condition on the minimal closure
it may exist some non elementary conditions:
If X ! A is a non elementary condition
and if X+ = U, then X is a superkey.
37
CSC343
Remark (end):
R(A,B,C,D)
F = f AB ! C, B ! D, D ! B, B ! A g
F is not a minimal closure, why?
(R, F ) is in 3NF, why?
38
CSC343
Lossless-Join Decomposition
Example
R ( A B C )
a b c
a b a
c b d
and F = f A ! B g
We decompose in:
R1 ( A B ) R2 ( B C )
a b b c
c b b a
b d
39
CSC343
R1 = (R) A;B
R2 = (R)
R’ = R1 1 R2 6= R:
B;C
R’ ( A B C )
a b c
a b a
a b d
c b c
c b a
c b d
The decomposition of R in R1 and R2 losses information.
The join creates tuples that do not exist in R.
40
CSC343
Now we decompose R’ in:
R1’( A B ) R2’ ( A C )
a b a c
c b a a
c d
R’’ = R1’ 1 R2’ = R’:
R’’ ( A B C )
a b c
a b a
c b d
this decomposition is lossless-join.
The condition is that after the join we found the same information than before the
decomposition.
41
CSC343
Definition A decomposition of R in R1, R2, ..., Rk with regard to a set of FD F is

lossless-join iff for every instance r of the schema R that satisfiesF , we have:
r = 1(r)1 2(r). . . 1 (r)

R R Rk
Theorem:
If (R1, R2) is decomposition of R and F a set of FD, then (R1, R2) is lossless-join
w.r.t. F , iff:
R1 \ R2 ! R1 - R2
or
R1 \ R2 ! R2 - R1
belongs to F + .
42
CSC343
Examples
R (A, B, C)
F =fA!Bg
1) R1(A, B), R2(B, C)
AB \ BC = B
AB - BC = A
BC - AB = C
The FD B ! A does not exist, nor B ! C in F +
) The decomposition loses information.
43
CSC343
2) R1(A, B), R3(A, C)

AB \ AC = A
AB - AC = B
A ! B is F (F +).
) the decomposition is lossless join.
44
CSC343
Dependency Preserving Decomposition
Definitions
1. Projection of a set of FD om Z U
(F ) = fX ! Y 2 F + j XY Z g
Z
Example: R(A,B,C,D), F = f AB ! C, C ! A, A ! D g
(F ) = f AB ! C, C ! A g
ABC
2. Decomposition which preserves the FDs of F

Let = (R1,...,Rk) be a decomposition, and F a set of FD.
preserves the FDs of F , if we can find again every FDs of F + from the union G of
all FDs projected from F in 1(F ), . . . , (F ):
R Rk
G+ = F +
45
CSC343
Examples
R(A,B,C,D)
F = f AB ! C, C ! A, A ! D g
= (ABC, BD) do not preserve the FDs of F
= (ABC, AD) preserves the FDs of F
R(A,B,C)
F = f A ! B, B ! A, A ! C g
= (AB, BC) preserve the FDs of F
46
CSC343
R(A, B, C, D)
F = f A ! B, B ! C, AB ! D g
The decomposition
R1(AC)
R2(AB)
R3(CD)
do not preserve the FDs of F . Why?
47
CSC343
Decomposition of a relation in 3NF
Given a schema (R, F) not in 3NF, i.e. with some anomalies, we want to find a
decomposition of R:
1. with 3NF relations;

2. lossless-join;
3. preserve the FDs of F
48
CSC343
Remark:
a lossless-join decomposition do not necessarily preserve the FDs and inversely.

the result does not necessarily give relations in 3NF.
Theorem: Every 1NF relation has a decomposition in 3NF relations which are
lossless-join and preserve the functional dependencies.
49
CSC343
Algorithm
We assume that F is a minimal closure.
1. For each X ! A 2 F , create a relation of schema (XA).

2. If no keys is contained in one a the schema created during the first step, add a relation of
schema Y where Y is one key.
3. If after the first step, it exist one relation R1 with a schema (X1A1) contained in a
schema (X2A2) of another relation R2, delete the relation R1.
4. Replace the relations (XA1),..., (XA ) (corresponding to FD with the same left
k
member) by a unique relation: (XA1...A ). k
50
CSC343
Examples
1) R(A,B,C,D)
F = f AB ! C, B ! D, C ! A g
Keys: AB, BC
Step 1: R1(ABC) R2(BD) R3(CA)
Step 2: No need to create a new relation:
the key AB belongs to R1
Step 3: Delete R3: CA ABC
Good decomposition: R1(ABC) R2(BD)
We can check that R1 et R2 are in 3NF.
51
CSC343
2) R(A,B,C,D,E)
F = f AB ! C, C ! D, C ! A g Keys: ABE, BCE
Step 1: R1(ABC) R2(CD) R3(CA)
Step 2: We add a relation of schema for the key ABE: R4(ABE)
Good decomposition: R1(ABC) R2(CD) R4(ABE)
Other solutions:
Step 4: We replace R2 and R3 from step 1 by the relation of schema (CAD)
Other good decomposition: R1(ABC) R2’(CAD) R4(ABE)
What happened if we have chosen the key CBE?
52
CSC343
3) R(A,B,C,D)
F = f AB ! C, C ! D, C ! A, AB ! D g
Keys: BA, BC
The relation is not in 3NF. Why?
Step 1: R1(ABC) R2(CD) R3(CA) R4(ABD)
Step 2: we do not add the relation: key AB R1(ABC)
Step 4: We replace R1 et R4 by R5(ABCD) ) we can delete R2.
Decomposition: R5(ABCD)
This decomposition is not in 3NF. Where is the problem?
53
CSC343
Boyce-Codd Normal Form (BCNF)
Some anomalies still exist in 3NF.
Example: Post(City,Street,ZipCode), F = f VC ! Z, V ! C g
Keys: CS, SZ
Post ( City Street ZipCode )

Toronto Queen M4F3G4
Toronto King M4F3G4
) Redondancy between the zipcode and the city.
54
CSC343
Definition: A relation is in Boyce-Codd Normal Form (BCNF), if for every functional

dependency of F , the left member id a superkey.
Interest: We eliminate all anomalies.
Remark: Every BCNF relation is in 3NF.
Unfortunately, it does not always exist a decomposition in BCNF which also:
is lossless-join
preserves the FD.
55
CSC343
Post Example
Post(City,Street,ZipCode), F = f CS ! Z, Z ! C g
Keys: CS, SZ
R is in 3NF but not in BCNF (in Z ! C, Z is not a key)
Post ( City Street ZipCode )

Toronto Queen M4F3G4
Toronto King M4F3G4
56
CSC343
The decomposition R1(City,ZipCode), R2(Street,ZipCode) avoids the

redondancy City,ZipCode, it is lossless-join, but does not preserve the functional
dependency CS ! Z
R1 ( City ZipCode ) R2 ( Strret Code )

Toronto M Queen M
Montreal T Queen T
The insertion of Toronto Queen M4, i.e. Toronto M4 and King M4 respects Z !
C but do not respect anymore CS ! Z
R1 ( Ville Code ) R2 ( Street Code )

Toronto M Queen M
Montreal T Queen T
Toronto M4 Queen M4
57
CSC343
Decomposition Algorithm
We assume that R(U) is a relational schema and F is a minimal cover.
1. Pick a FD X ! Y not verifying BCNF

2. Partition R into R1(X Y) and R2(X (U-Y))
3. If R1 is not in BCNF start the algorithm with R1 in input
4. If R2 is not in BCNF start the algorithm with R2 in input
58
CSC343
Example 1
R(BOSQID), F = f IS ! Q, B ! O, I ! B, S ! D g Candidate Key: IS
Pick B ! O:
R1 (BO)
R2 (BSQID)
Now we decompose R2 using S ! D
R3 (SD)
R4 (BSQI)
Now we decompose R4 using I ! B
R5(IB)
R6(ISQ)
59
CSC343
Set of relational schema which are in BCNF from R and F :
R1 (BO) F1 = f B ! O g
R3 (SD) F3 = f S ! D g
R5 (IB) F5 = f I ! B g
R6 (ISQ) F6 = f IS ! Q g
60
CSC343
Example 2
R(ABCDEF), F = f A ! BC, D ! AF g Candidate Key: DE
R1(ABC) F1 = f A ! BC g
R2(ADF) F2 = f D ! AF g
R3(DE) F3 = ;
61
CSC343
Example 3
R(ABC), F = f AB ! C, C ! A g Candidate Key: AB,CB
R1(AC) F1 = f C ! A g
R@(DE) F2 = ;
62
CSC343
63
CSC343
4NF
CSC343
Definition
Functional dependencies rule out certain tuples from appearing in a relation: if A ! B,

then we cannot have two tuples with the same A value but different B values.
Multivalued dependencies do not rule out of the existence of certain tuples. Instead they
require that other tuples of a certain form be present in the relation.
Let R(U, F ) be a relational schema and A,B U. The multivalued dependency:
A B
holds in R if any legal relation r(R), for all pairs of tuples t1 and t2 in r such that t1[A] =
t2[A], there exist tuples t3 and t4 in r such that:
– t1[A] = t2[A] = t3[A] = t4[A]
– t1[B] = t3[B]
– t2[R-B] = t3[R-B]
– t2[B] = t4[B]
– t4[R-B] = t1[R-B]
65
CSC343
Example
name address car

Tom North Rd. Toyota
Tom Oak St. Honda
Tom North Rd. Honda
Tom Oak St. Toyota
MV D = fname address; name carg
66
CSC343
Example
name street city title year

C. Fisher 123 Maple St. Hollywood Star Wars 1977
C. Fisher 5 Locust Ln. Malibu Star Wars 1977
C. Fisher 123 Maple St. Hollywood Empire Strike Back 1980
C. Fisher 5 Locust Ln. Malibu Empire Strike Back 1980
C. Fisher 123 Maple St. Hollywood Return of the Jedi 1983
C. Fisher 5 Locust Ln. Malibu Return of the Jedi 1983
F = fname street title year ! cityg

MV D = fname street city g
67
CSC343
Theory of Multivalued Dependencies
We will need to compute all the multivalued dependencies that are logically implied by a
given set of multivalued dependencies.
Let D denote a set of functional and multivalued dependencies.

The closure of D+ is the set of all functional and multivalued dependencies logically
implied by D.
We can compute from D+ using the formal definitions, but it is easier to use a set of
inference rules.
68
CSC343
Theory of Multivalued Dependencies
The following set of inference rules is sound and complete. The first three rules are
Armstrong’s axioms.
Reflexivity rule: if X is a set of attributes and Y X, then X ! Y holds.

Augmentation rule: X ! Y j= XZ ! Y Z .
Transitivity rule: if X ! Y holds, and Y ! Z holds, then X ! Z holds.
Complementation rule: if X Y holds, then X (R B A) holds.
Multivalued augmentation rule: if X Y holds, and Z R and T Z , then
XZ T B holds.
Multivalued transitivity rule: if X Y holds, and Y Z holds, then X Z Y
holds.
Replication rule: if X ! Y holds, then X Y .
Coalescence rule: if X ! Y holds, and Z Y , and there is a T such thatT R and
T \ Y = ; and T ! Z , then A ! Z holds.
69
CSC343
An example of multivalued transitivity rule is as follows. If we have R(A,B,C,D) and

A B and B B; C . Thus we have A C , where C = B; C B
An example of coalescance rule is as follows. If we have R(A,B,C,D) and A B; C ) and
D ! B , then we have A ! B
A MVD X1X2:::X n Y1Y2:::Ym for a relation R is nontrivial if:
1. None of the Y’s is among the A’s.

2. Not all the attributes of R are among the A’s and B’s.
70
CSC343
Other Axioms
Multivalued Union rule: if X Y holds and X Z holds, then X Y; Z holds.

Intersection rule: If X Y holds and X Z holds, then X Y \ Z holds.
Difference rule: If X Y holds, and X Z , then X Y Z holds and X Z Y
holds.
71
CSC343
Fourth Normal Form (4NF)
We saw that a BCNF schema was not an ideal design as it suffered from repetition of
information.
We can use the given multivalued dependencies to improve the database design by
decomposing it into fourth normal form.
A relation schema R is in 4NF with respect to a set D of functional and multivalued
dependencies if for all multivalued dependencies in D+ of the form X Y , where
X R and Y R :
–X Y is a non trivial multivalued dependency; and
– X is a superkey for schema R.
A database design is in 4NF if each member of the set of relation schemas is in 4NF.
The definition of 4NF differs from the BCNF definition only in the use of multivalued
dependencies:
– Every 4NF schema is also in BCNF.
72
CSC343
Summarize
Property 3NF BCNF 4NF

Eliminates redundancy due to FD’s Most Yes Yes
Eliminates redundancy due to MVD’s No No Yes
hline Preserves FD’s Yes Maybe Maybe
Preserves MVD’s Maybe Maybe Maybe
73
CSC343

Functional Dependencies

Uploaded by

Copyright:

Available Formats

Functional Dependencies

Uploaded by

Document Information

Original Description:

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

Functional Dependencies

Uploaded by

Copyright:

Available Formats

FUNCTIONAL DEPENDENCIES

Examples: PERSON Relation

PERSON SIN NAME CITY

 “If I know the sin number I know the name”

COURSE NAME PROF STUDENT

 “A Course has only one Professor”

PERSON (SIN, LastName, FirstName, Address)

 If we know the SIN value, we know all the attributes.

SUPP (SNAME, PNAME, QUANT, PRICE)

SNAME PNAME ! QUANT

Quant SUPPLYING Price

ADDRESS (STREET, NUMBER, CITY, ZIPCODE)

 Several zipcode for a city: Toronto = M5B3F4, M5S2E4, . . .

Number Street City ! Zipcode

Finding a Key: Example

Functional Dependencies Properties

Closing set of a functional dependencies set

2) R(Course, Prof, Hour, Room, Student, Mark)

1. We can infer from G the same FD than from F : G + = F +

Functional Minimal Closure

ACD ! B is not elementary, why?

ACD ! B is not elemenaty, why?

C ! A j= CCD ! ACD by augmentation

CCD ! ACD j= CD! ACD (A ! A by reflexivity and union)

CD ! ACD and ACD ! B j= CD ! B by transitivity

By pseudotransitivity, f C ! A, ACD ! Bg j= CCD ! B j= CD ! B.

Concretely, the minimal closure

 is either directly given by the real world analysis.

New Definition for a SuperKey

Let R(U) a relational schema, F a set of functional dependencies and X

But show that F j= X ! A is easy and fast.

Calculation of a key of a relation

Let R(U) a relational schema and X  U a set of attributes.

 We define X + as a subset of attributes A such that F j= X ! A

1. We search for one X such that X + = U ) X is a superkey

Step 1: We start from X

For every Y ! A, such that Y  X 1, Y

We stop when we do not find new FDs anymore.

Let the schema S1:

and the set of DF: F = fSNAME ! FADDRESS, (SNAME PNAME) ! PRICEg

Suppose now that we replace S1 by S2:

R(SNAME, SADDRESS, PNAME, PRICE)

R(SNAME, SADDRESS, PNAME, PRICE)

F = fSNAME ! SADDRESS, (SNAME PNAME) ! PRICEg

What is the key of R?

SMITH Toronto COMPUTER 1000

1) REDONDANCY: the address for one supplier appears several times.

=) THE INITIAL SCHEMA S1 IS “BETTER”

It exists several types of constraints on the tuples:

1. dependencies (functional, multivalued, ...)

Qualities for a good schema

1. Avoid anomalies =) decomposition

Relation in first normal form (1NF)

 All attributes are atomic (elementary)

MARKS ( COURSE STUDENT MARK )

MARKS ( COURS PERF (STUDENT MARK) )

Relation in 3rd Normal Form

 2nd Normal Form:

3NF: First Definition

“If I know the sin number I know the name”

“A Course has only one Professor”

If we know the SIN value, we know all the attributes.

Several zipcode for a city: Toronto = M5B3F4, M5S2E4, . . .

is either directly given by the real world analysis.

Let R(U) a relational schema and X U a set of attributes.

We define X + as a subset of attributes A such that F j= X ! A

For every Y ! A, such that Y X 1, Y

All attributes are atomic (elementary)

2nd Normal Form:

r = 1(r)1 2(r). . . 1 (r)

a lossless-join decomposition do not necessarily preserve the FDs and inversely.

Functional dependencies rule out certain tuples from appearing in a relation: if A ! B,

MV D = fname address; name carg

Let D denote a set of functional and multivalued dependencies.

Reflexivity rule: if X is a set of attributes and Y X, then X ! Y holds.

Multivalued Union rule: if X Y holds and X Z holds, then X Y; Z holds.