Chapter 7: Functional Dependences and Normalization

Outline:

7.1 Problems of relational design

 Informal Guidelines Design for Relational Databases
o Semantics of the relation attributes.
o Redundant values in tuples and update anomalies.
o Null values in tuples.
o Generating spurious tuples.

7.2 Functional dependences.

 Definitions.
 Trivial VS. Nontrivial FDs
 Inference Rules for Functional Dependencies.
o Armstrong’s Axioms
o Additional Inference rules
 Closure of attribute sets, X+

7.3 Normalization.
 Definitions.
 First Normal Form (1NF)
 Second Normal Form (2NF)
 Third Normal Form (3NF)
 Boyce-Codd Normal Form (BCNF)

7.1 Problems of relational design

What are the criteria for “good” relational schemas?

 We first discuss informal guidelines for good relational design
 Then we discuss formal concepts of functional dependencies and normal
forms
o 1NF (First Normal Form)
o 2NF (Second Normal Form)
o 3NF (Third Normal Form)
o BCNF (Boyce-Codd Normal Form)

1
Informal Guidelines Design for Relational Databases
 There are four informal measure of quality for relation schema design:
1. Semantics of the relation attributes.
2. Redundant values in tuples and update anomalies.
3. Null values in tuples.
4. The possibility of generating spurious tuples.
 These measures are not always independent of one another.

1. Semantics of the relation attributes

o Design a relation schema so that it is easy to explain its meaning.
o Do not combine attributes from multiple entity sets and relationship sets
into a single relation.

Guideline1:
Map each entity set to a relation & each relationship set to a relation.
o Don’t mix these two
o Only use foreign keys to refer to other entities as opposed to
duplicating attributes in other relations.

Example: Consider the following “EMP_DEPT” relation.

EMP_DEPT

EName SSN BDate Address DNumber DName DMGRSSN

It is good design?
o The attributes semantic of the EMP_DEPT relation are not clear.
o It is poor design because they violate guideline 1 by mixing attributes from
distinct real-world entities.
o Solution: decompose the “EMP_DEPT” relation into two relations and foreign
key.
EMP DEPT

EName SSN BDate Address DNumber DNumber DName DMGRSSN

2. Redundant value in tuples and update anomalies

o Data redundancy refers to copying of the same data in the same table or
multiple tables.

2
 Example: Consider the following “EMP_DEPT” relation.

EMP_DEPT

EName SSN BDate Address DNumber DName DMGRSSN

John 12345 09-01-1965 Houston 5 Research 33455

Alicia 33455 08-08-1955 Bellaire 5 Research 33455

….

Redundancy

o There are problems that result from having such large table:

1. There is a waste of storage space due of data redundancy.

o In the previous example, the attribute values (DNumber, DName,
and DMGRSSN) are repeated for every employee who works for
the same department.

2. There are anomalies in updating such relation

o These can be classified into insertion anomalies, deletion
anomalies, and modification (update) anomalies.

A. Insertion Anomalies: there are two types:

 To insert a new employee tuple into EMP_DEPT,
we must include either the consistent attribute value
for the department that the employee works for, or
nulls (if the employee does not work for a
department as yet).

 It is difficult to insert a new department that has no

employees as yet in the EMP_DEPT relation. The
only way to do this is to place null values in the
attributes for employee. This cause a problem
because SSN is the primary key of EMP_DEPT, and
each tuple is supposed to represent an employee
entity not a department entity.

B. Deletion Anomalies:
 Deleting departments will delete their employees.

 If we delete an employee tuple that happens to

represent the last employee working for a particular

3
 department, the information concerning that
department is lost from the database.

C. Modification (Update) Anomalies:

 Updates causing the updating of too many rows. for
example, if we change the department name of
department 5 – we must update the tuples of all
employees who work in that department; otherwise,
the database will become inconsistent.

Guideline2:
Minimize the amount of data redundancy in the database.

3. Null values in tuples

o If some attributes do not apply to many tuples, then the occurrence of data
in those columns is sparse and storage space is wasted.

o Example: if only 10% of employees have individual office, there is little

justification for including the office attributes in the employee relation;
rather, a relation EMP_OFFICE (SSN, OfficeNo, OfficeLoc, OfficeTelNo)
can be created to include tuples for only the employees with individual
offices.

o Queries involving tables having null values might have difficulty dealing
with the null values, namely joins and aggregates.

Guideline3:
o Try to minimize the occurrences of NULL values in your design.
o Attributes that are NULL frequently could be placed in separate
relations (with the primary keys).

4. Spurious Tuples

o Decomposition is the process of breaking down larger tables into smaller

tables. If the resulting tables are then joined, we should get the original
large table. If we do, the decomposition is called lossless decomposition. If
we do not get the original table, the decomposition is called lossy
decomposition.

4
 o Example: Consider the following table

Employee
EID EName TelNo SDate
12345 Kim 555-1234 01/23/2004
54321 Kim 555-5432 05/03/2001

We break up the Employee table into two smaller tables

Emp1
EID EName
12345 Kim
54321 Kim

Emp2
EName TelNo SDate
Kim 555-1234 01/23/2004
Kim 555-5432 05/03/2001

Then we join these two tables back together

EID EName TelNo SDate

12345 Kim 555-1234 01/23/2004
12345 Kim 555-5432 05/03/2001 Spurious
54321 Kim 555-1234 01/23/2004 Tuples
54321 Kim 555-5432 05/03/2001

This is not the same table. Thus, this is a lossy decomposition.

o To do lossless decomposition, you must decompose on a key (a key

existing in the new tables).

Guideline4:
o The relations should be designed to satisfy the lossless join
condition.
o No spurious tuples should be generated by doing a natural-join of
any relations.

5
7.2 Functional Dependencies

 Functional Dependencies FDs are:

o Tools for analysis.
o Used to specify formal measures of the "goodness" of relational
designs. And keys are used to define normal forms for relations
o Constraints that are derived from the meaning and interrelationships of
the data attributes.

Definitions

 FD is a constraint between two sets of attributes from the database. Require

that the value for a certain set of attributes determines uniquely the value for
another set of attributes.
 A set of attributes X functionally determines a set of attributes Y if the value
of X determines a unique value for Y.
 X→Y means Y is functionally dependent on X or X functionally determines
Y.
 X→Y holds if whenever two tuples have the same value for X, they must have
the same value for Y.
o For any two tuples t1 and t2 in any relation instance r(R): if t1[X] =
t2[X], then t1[Y] = t2[Y].
 If X→Y holds in r, this does not say whether or not Y→X holds in r.
 Example: consider r(A,B) with the following instances of r:

A B
1 4
2 5
1 7
On these instances, A→B does not hold, but B→A does hold.

 Example: Consider r(X,Y,Z) with the following instances of r:

X Y Z
X1 Y1 Z1
X2 Y1 Z1
X3 Y2 Z3
X4 Y3 Z2

X→Y but Y not determine X

X→Z but Z not determine X
Y→Z and Z→Y

o Note: if X is a candidate key of r – this implies that X→ r. because the

key constraint implies that no two tuples in any legal state r(R) will
have the same value of X.

6
  K is a superkey for relation schema r if and only if K→r (K determines all
attributes of r. all attributes of relation occur on the right-hand side of the FD).
 K is a candidate key for r if and only if
 K→r, and
 No α K, α→r (no subset of K is a superkey)

 Example: Consider r(A,B,C) with the following instances of r:

A B C
A1 B1 C1
A2 B1 C1
A1 B2 C1
A4 B3 C2
Is {A,B} candidate key?

 Example: Consider the relation EMP (SSN, Ename, Deptno). Suppose EMP
is used to represent many-to-one relationship from EMP to DEPT, where SSN
is a key in the EMP relation and Deptno is a key in the DEPT relation.
o Are the following FDs holds in EMP relation?
1. {SSN}→{Deptno}
2. { Deptno}→{SSN}
o What about many-to-many relationship and one-to-one relationship?

 FD is a property of the semantics or meaning of the attributes.

o Examples:
{SSN} → {Ename}
{PNumber} → {Pname, Plocation}
{SSN, Pnumber} → {Hours}

 Relation extensions r(R) that satisfy the FD constraints are called legal
extensions (or legal relation states) of R.
 A FD is a property of the relation schema (intension) R, not of a particular
legal relation state (extension) r of R. hence, an FD cannot be inferred
automatically from a given relation extension r but must be defined explicitly
by someone who knows the semantics of the attributes of R.
 Example: Consider the following relation

Teacher Course Text

T1 DS X
T1 C++ Y
T2 Java Z
We can infer from the relation state (extension) that Course →
Text and Text → Course. After insert the following tuple
T3 DS W
The Text → Course but Course not determine Text. After that,
when insert the following tuple
T4 Advanced Z
Java

7
 The Text not determine the Course

 FD constraint must hold at all times.

Trivial VS. Nontrivial FDs

 A FD is trivial iff RHS (dependent) is a subset of the LHS (determinant).

X→Y is trivial if Y  X
 Examples:
{SSN}→{SSN}
{SSN,Ename}→{Ename}
 A FD is nontrivial if RHS is not a subset of the LHS.
X→Y is nontrivial if Y  X
 Example:
{SSN}→{Ename}

Inference Rules for Functional Dependencies

 Given a set of FDs F, we can infer additional FDs that hold whenever the FDs
in F hold
 The set of all dependences is called the closure of F and is denoted by F+.

A. Armstrong’s Axioms
 We can find all of F+ by applying Armstrong’s axioms.

1. IR1: Reflexivity rule

If Y  X, then X→Y
Ex: {CNo,CName,BDate} → {CName,BDate}

2. IR2: Augmentation rule

X→Y ╞ XZ→YZ
Ex: {CNo}→{CName} ╞ {CNo,BDate}→{CName,BDate}

3. IR3: Transitive rule

{X→Y, Y→Z} ╞ {X→Z}
Ex: {CNo→BDate, BDate→Age_Category} ╞
{CNo→Age_Category}

 These rules are:

 Sound: generate only functional dependencies that
actually hold. And
 Complete: generate all functional dependencies that
hold.

 Example: Consider R(X,Y,Z) with the following FDs

F={X→Y,
X→Z}
We can infer the following FDs

8
 1. X→XY (Augmentation by X)
2. YX→YZ (Augmentation by Y)
3. X→YZ (Transitivity)

 Example: Consider R(A,B,C,G,H,I) with the following FDs

F={A→B,
A→C,
CG→H,
CG→I,
B→H}
Proof the following FDs
1. A→H
By transitivity from A→B and B→H

A→B
Transitive B→H
B→H

2. AG→I
By augmentation A→C with G, to get
AG→CG and then transitivity with CG→I

CG→I
Transitive AG→I
A→C Aug AG→CG

3. CG→HI
By augmentation CG→I with CG, to get
CG→CGI, then by augmentation CG→H
with I, to get CGI→HI, and then by
transitivity CG→HI

Aug
CG→I CG→CGI
Aug Transitive CG→HI
CG→H CGI→HI

B. Additional Inference rules

 Can be derived from Armstrong’s Axioms.

 We can further simplify manual computation of F+ by using the
following additional rules.

9
 4. IR4: Decomposition rule
{X→YZ}╞ {X→Y and X→Z}

By reflexivity YZ→Y
Transitive X→Y
X→YZ

By reflexivity YZ→Z
Transitive X→Z
X→YZ

5. IR5: Union rule

{X→Y, X→Z} ╞ {X→YZ}

Aug
X→Y X→XY
Transitive X→YZ
Aug
X→Z XY→YZ

6. IR6: Composition rule

{X→Y, W→Z} ╞ {XW→YZ}

Aug
X→Y XW→YW
Transitive XW→YZ
Aug
W→Z WY→YZ

7. IR7: Psuedotransitive rule

{X→Y, WY→Z} ╞ {WX→Z}

WY→Z
Transitive WX→Z
Aug
X→Y XW→YW

Closure of attribute sets, X+

 F+ can grow quite large, as we keep applying rules to find more FDs.

10
  Sometimes we want to find all of F+, and other times we just want to find part
of it.
 We are often interested in finding the part that tells us whether or not some
subset of attributes x is a superkey for R.
 If you can uniquely determine all attributes in R by some subset of attributes
X, then X is a superkey for R.
 The closure of X under F, denoted X+, is the subset of attributes that are
uniquely determined by X under F.

 Definition:
o Given a schema R, a set X of attributes in R, and a set F of FDs that
hold for R, then the set of all attributes of R that are functionally
dependent on X is called closure of X under F (X+)

 Algorithm: Determining X+, the closure of X under F.

X+ := X;
Repeat
oldX+ := X+ ;
for each functional dependency Y→Z in F do
if Y  X+ then X+ := X+ U Z;
Until (X+ = oldX+);

 Example: Consider R(A,B,C,D,E,F) with the following FDs

F={A→BC,
B→CF,
B→E,
CD→EF}
Find {AB} +

oldX+ X+
AB AB
ABC
ABCF
ABCEF
ABCEF

{AB}+ = {A B C E F}
{AB} is not a superkey because it is not determine R

 Example: Consider R(A,B,C,G,H,I) with the following FDs

F={A→B,
CG→H,
CG→I,
A→C,
B→H}

11
 Is {AG} a candidate key?

oldX+ X+
AG AG
ABG
ABCG
ABCGH
ABCGH
ABCGHI
ABCGHI

{AG}+ = {A B C G H I}
{AG} is a superkey

{AG} is a candidate key, because no subset of AG is a

superkey. ({A} + = {A B C H} and {G}+ = {G})

 Example: Consider R(A,B,C,D,E) with the following FDs

F={AB→C,
A→D,
D→E,
AC→B}
Is {AB} a candidate key?

7.3 Normalization

 Formal process for designing relational database schema.

 Start from larger tables and decompose them until each smaller table follows a
"good" form.
 The normalization process takes a database schema through a series of steps
that reduce at each step the amount of data redundancy and the number of
anomalies.
 In normalization theory, there are multiple levels of normal forms, denoted as
NF1, NF2, etc. which represent each step in the normalization process.
 We assume that a set of functional dependencies is given for each relation, and
that each relation has a designated key. The FDs and Keys are used to tests for
normal form.
 The normalization process, as first proposed by Codd (1972), takes a relation
schema through a series of tests to certify whether it satisfies a certain normal
form.
 1NF, 2NF, 3NF, and BCNF based on the functional dependencies and keys.

Definitions:
 A superkey of a relation schema R = {A1, A2, ...., An} is a set of attributes S
subset-of R with the property that no two tuples t1 and t2 in any legal relation
state r of R will have t1[S] = t2[S].

12
  A key K is a superkey with the additional property that removal of any
attribute from K will cause K not to be a superkey any more.
 If a relation schema has more than one key, each is called a candidate key.
o One of the candidate keys is arbitrarily designated to be the primary
key, and the others are called alternate keys.
 A Prime attribute must be a member of some candidate key.
 A Nonprime attribute is not a prime attribute—that is, it is not a member of
any candidate key.

First Normal Form (1NF)

 Disallows having a set of values (multivalue attribute), a tuple of values

(composite attribute), or a combination of both (complex attribute) as an
attribute value for a single tuple.
 The only attribute values permitted by 1NF are single atomic (or indivisible)
values.

 Example1: (Multivalue attribute)

Consider a relation DEPARTMENT, which has a multivalue attribute

DLocation. Each department can have multiple locations.
DEPARTMENT (Dname, Dnumber, DMGRSSN, {Dlocation})

Dname Dnumber DMGRSSN Dlocation

There are two ways we can look at the Dlocations attribute:

1. The domain of Dlocation contains atomic values. In this case,
Dlocation is not functionally dependent on Dnumber.

Dname Dnumber DMGRSSN Dlocation

Research 5 33455 LocX
Research 5 33455 LocY
Research 5 33455 LocZ
Administration 4 98765 LocW
Headquarters 1 88677 LocZ

2. The domain of Dlocation sets of values and hence is nonatomic

in this case, Dnumber→Dlocation.

Dname Dnumber DMGRSSN Dlocation

Research 5 33455 {LocX, LocY,
LocZ}
Administration 4 98765 LocW
Headquarters 1 88677 LocZ

13
 There are three main techniques to achieve first normal form for such a
relation:
I. First Solution: Decompose the non 1NF relation into two 1NF
relations, by remove the attribute Dlocation that violate 1NF
and place it in a separate relation Dept_Location. The primary
key of this relation is the combination (Dnumber, Dlocation).

Dname Dnumber DMGRSSN

Research 5 33455
Administration 4 98765
Headquarters 1 88677

Dnumber Dlocation
5 LocX
5 LocY
5 LocZ
4 LocW
1 LocZ

This solution is the best because no redundancy in the

relations and no limit placed on a maximum number of
location values.

II. Second solution: Expand the key so that there will be a separate
tuple in the original DEPARTMENT relation for each location
of a department.
The primary key is (Dnumber, Dlocation)

Dname Dnumber DMGRSSN Dlocation

Research 5 33455 LocX
Research 5 33455 LocY
Research 5 33455 LocZ
Administration 4 98765 LocW
Headquarters 1 88677 LocZ

-ve: introduce redundancy in the relation.

III. Third solution: If a maximum number of values is known for

the attribute – for example, if it is known that at most three
locations can exist for a department. Replace the Dlocation
attribute by three atomic attributes: Dloc1, Dloc2 and Dloc3

Dname Dnumber DMGRSSN Dloc1 Dloc2 Dloc3

Research 5 33455 LocX LocY LocZ
Administration 4 98765 LocW NULL NULL

14
 Headquarters 1 88677 LocZ NULL NULL

-ve: introduce Null values in the relation if most

departments have fewer than 3 locations. This solution
breaks guideline #3 by introducing too many nulls and thus
wasting space.
 Example2: (Composite attribute)
Consider a relation EMPLOYEE, which has a composite attribute Name.
Each Name have first and last name. EMPLOYEE (SSN, Name (FName,
LName), Address)

SSN Name Address

FName LName
2314 Adam James X
6543 Sali John Y
7642 Smith King Z

The EMPLOYEE relation is not in 1NF, because the Name attribute not
atomic attribute

Solution, separate each component in the individual column.

SSN FName LName Address

2314 Adam James X
6543 Sali John Y
7642 Smith King Z

 Example3: (Complex attribute (nested relations))

Consider a relation EMP_PROJ, which has a complex attribute Projs.
Which consist of Pnumber and Hours.
EMP_PROJ (SSN, Ename, Projs {(Pnumber, Hours)})

SSN Ename PNumber Hours

2314 Adam 1 32.5
2 7.5
6543 Sali 3 40.5
7642 Smith 2 10.0
3 10.0
10 15.0
20 12.0

SSN is the primary key of the EMP_PROJ relation and Pnumber is the
partial key of the nested relation; that is, within each tuple, the nested
relation must have unique values of Pnumber.

15
 To normalize this into 1NF, we remove the nested relation attributes into
a new relation and propagate the primary key into it; the primary key of
the new relation will combine the partial key with the primary key of the
original relation.

Solution:

SSN Ename

SSN PNumber Hours

Second Normal Form (2NF)

 2NF is based on the concept of full functional dependency.

 X→Y is a fully functional dependency, if removal of any attribute A from X
means that the dependency does not hold any more; that is, for any attribute A
X, (X-{A}) does not functionally determine Y.
 X→Y is a partial dependency if some attribute A X can be removed from X
and the dependency still holds.
 Example:
{SSN, Pnumber}→{Hours} is a full dependency, because neither
{SSN→Hours} nor {Pnumber→Hours} holds.

{SSN, Pnumber}→{Ename} is a partial dependency, because

{SSN→Ename} holds.

A. 2NF based on the primary key

o A relation schema R is in second normal form (2NF) if every non-

prime attribute A in R is fully functionally dependent on the primary
key
 If the primary key contains a single attribute, the relation in the
2NF.
o Example: Consider the following relation. Determine if the relation in
2NF, if not normalize the relation into 2NF.

EMP_PROJ
SSN Pnumber Hours Ename Pname Plocation
FD1

FD2
FD3

 There are two problems in the previous relation:

16
 A. The semantic of the attributes are not clear.
B. Redundant data (update anomalies)
 EMP_PROJ relation is in 1NF but is not in 2NF,
because the nonprime attribute Ename violate 2NF
(FD2); as do the nonprime attributes Pname and
Plocation (FD3).
 FD2 and FD3 make Ename, Pname and Plocation
partially dependent on the primary key {SSN,
Pnumber} of EMP_PROJ
 Solution: separate each FD violate 2NF into new
relation as follows:

SSN PNumber Hours SSN Ename

FD1 FD2

Pnumber Pname Plocation

FD3

o After decomposition:
 No redundant values
 No spurious tuples through join

o Example: Consider the following relation. Determine if the relation in

2NF, if not normalize the relation into 2NF.

Dept Item Price

D1 T3 50
D1 T2 40
D2 T5 40
D2 T2 40

FD1: {Dept, Item}→{Price}

FD2: {Item}→{Price}
The relation is in 1NF but it is not in 2NF (FD2)

Solution:

Dept Item Item Price

D1 T3 T3 50
D1 T2 T2 40
D2 T5 T5 40
D2 T2

17
 B. General definition of 2NF

o A relation schema R is in 2NF if every nonprime attribute A in R is

fully dependent on any key of R.
o Example: Consider the following relation. It is in 2NF or not?

A B C D E F
FD1
FD2

The relation in 1NF but it is not in 2NF (FD2)

Solution:

A B C D
FD1

D E F
FD2

Third Normal Form (2NF)

 3NF is based on the concepts of transitive dependency.

 X→Y in a relation schema R is a transitive dependency if there is a set of
attributes Z that is neither a candidate key nor a subset of any key of R, and
both X→Z and Z→Y hold.

A. 3NF based on the primary key

o A relation schema R is in third normal form (3NF) if it is in 2NF and

no non-prime attribute A in R is transitively dependent on the primary
key.
o Example: Consider the following relation. Determine if the relation in
3NF, if not normalize the relation into 3NF.

EMP_DEPT
Ename SSN BDate Address Dnumber Dname DMGRSSN
FD1
FD2

{SSN→DMGRSSN} is transitive dependent through

Dnumber (SSN→Dnumber, Dnumber→DMGRSSN),
and Dnumber is neither a key itself nor a subset of the
key of EMP_DEPT.

18
 Solution:
ED1
Ename SSN BDate Address Dnumber
FD1

ED2
Dnumber Dname DMGRSSN
FD2

Note: A natural join operation on ED1 and ED2 will

recover the original relation EMP_DEPT without
generating spurious tuples.

o Example: Consider the following relation. Determine if the relation in

3NF, if not normalize the relation into 3NF.

Item Category Discount

T3 A 10
T2 B 2
T5 C 12
T7 B 2

FD1: {Item}→{Category}
FD2: {Item}→{Discount}
FD3: {Category}→{Discount}
The relation is in 1NF and in 2NF but it is not in 3NF (FD3)

Solution:

Item Category Category Discount

T3 A A 10
T2 B B 2
T5 C C 12
T7 B

B. General definition of 3NF

o A relation schema R is in 3NF if, whenever a nontrivial functional

dependency X→A holds in R, either
1. X is a superkey of R.
Or
2. A is a prime attribute of R.

19
 o Example: Consider the following relation

Deptno DName Location

10 R LocX
20 A LocY
30 H LocZ

FD1: {Deptno}→{DName, Location}

FD2:{DName}→{Location}

The relation is in 3NF because DName is a superkey.

o Example: Consider the following relation

SSN Pnumber Hours

1234 101 10
5678 101 10
1234 103 20
4444 101 10
4444 103 30

FD1: {SSN, Pnumber}→{Hours}

FD2:{Hours}→{Pnumber}

The relation is in 3NF because Pnumber is a prime attribute.

 Example: Consider the following relation, it is in the 3NF? If not normalize

the relation into 3NF.
LOTS
Property_ID# County_Name Lot# Area Price Tax_Rate
FD1
FD2
FD3
FD4

The LOTS relation in 1NF but it is not in 2NF (FD3)

Decompose LOTS into two relations as follows

LOTS1 LOTS2
Property_ID# County_Name Lot# Area Price County_Name Tax_Rate

FD1 FD3
FD2
FD4

20
LOTS1 in 2NF but not in 3NF (FD4)
Decompose LOTS1 into two relations as follows

LOTS1A LOTS1B
Property_ID# County_Name Lot# Area Area Price

FD1 FD4
FD2

 Example: Consider the following relation, it is in the 3NF? If not normalize

the relation into 3NF.

BOOK_AUTH
Book Author Book List Publisher Author
Publisher
Title Name Type Price Date Affil
FD1
FD2
FD3

The BOOK_AUTH relation in 1NF but it is not in 2NF (FD1 and FD3)

Decompose BOOK_AUTH into three relations as follows

B_A1 B_A2 B_A3

Book Author Publisher Book Book List Author Author
Publisher
Title Name Date Title Type Price Name Affil

FD1
FD2

The B_A2 in 2NF but not in 3NF (FD2)

Decompose B_A2 into two relations as follows

B_A2_1 B_A2_2
Book Book Book List
Publisher
Title Type Type Price

FD1 FD2

Finally there are four relations (B_A1, B_A3, B_A2_1, B_A2_2)

21
Boyce_Codd Normal Form (BCNF)

 BCNF was proposed as a simpler form of 3NF, but it was found to be stricter
than 3NF, because every relation in BCNF is also in 3NF; however, a relation
in 3NF is not necessarily in BCNF.
 A relation schema R is in BCNF if whenever a nontrivial FD X→A holds in
R, then X is a superkey of R.
 The only difference between the definitions of BCNF and 3NF is the second
condition of 3NF, which allows A to be prime, is absent from BCNF.

 Example: Consider the following relation

A B C
FD1
FD2

a1 b1 c1
a1 b2 c3
a2 b2 c4
a3 b2 c3
a3 b3 c5

The relation is in 3NF but it is not in BCNF (FD2)

Solution: Decompose the relation into two relations

A C C B

Note: the decomposition lose the functional dependency FD1

 Example: Consider the following relation

Student Course Instructor

Jamal DB Randa
Hosam DB Sawsan
Hosam OS Ali
Hosam Theory Osama
Nada DB Randa
Nada OS Belal
Mohammad DB Anas
Omar DB Sawsan
FDs:
FD1: {Student, Course}→{Instructor}

22
 FD2: {Instructor}→{Course}

The relation is in 3NF, but it is not in BCNF (FD2)

Solution:
There are three possible decomposition of the relation
1. {student, instructor} and {student, course}
2. {course, instructor } and {course, student}
3. {instructor, course } and {instructor, student}
Out of the above three, only the 3rd decomposition will not generate
spurious tuples after join.

Student Instructor Instructor Course

Jamal Randa Randa DB
Hosam Sawsan Sawsan DB
Hosam Ali Ali OS
Hosam Osama Osama Theory
Nada Randa Belal OS
Nada Belal Anas DB
Mohammad Anas
Omar DB

 Example: Consider the following relation for Chocolate Factory

Product Prod
CustID Name Order# OrderDate Price Qty
Code Desc
C004 Adams, Anne Ord001 03/08/2000 P100 TopDeck 3.25 150
C004 Adams, Anne Ord001 03/08/2000 P300 KitKat 3.10 100
C003 Jones, Carol Ord002 05/08/2000 P200 BarOne 2.95 240
C002 Black, Roger Ord003 05/08/2000 P500 MilkyBar 3.20 370
C002 Black, Roger Ord003 05/08/2000 P400 Flake 3.40 120
C001 Smith, John Ord004 09/08/2000 P300 KitKat 3.10 280
C005 Rhodes, Sean Ord005 12/08/2000 P400 Flake 3.40 150
C002 Black, Roger Ord006 13/08/2000 P100 TopDeck 3.25 320
C001 Smith, John Ord007 15/08/2000 P500 MilkyBar 3.20 240

How would you split the previous relation to minimize data redundancy?

23