Chapter 8:

Relational Database
1
Design

D. NAGA JYOTHI CSE DEPT.

Chapter 8: Relational Database Design
2
 Features of Good Relational Design
 Atomic Domains and First Normal Form
 Decomposition Using Functional Dependencies
 Functional Dependency Theory
 Algorithms for Functional Dependencies
 Decomposition Using Multivalued Dependencies
 More Normal Form
 Database-Design Process
 Modeling Temporal Data

D. NAGA JYOTHI CSE DEPT.

3 Combine Schemas?

 Suppose we combine instructor and department into

inst_dept
 (No connection to relationship set inst_dept)
 Result is possible repetition of information

D. NAGA JYOTHI CSE DEPT.

 4 A Combined Schema
 Consider combining relations
Without
 sec_class(sec_id, Repetition
building, room_number) and
 section(course_id, sec_id, semester, year)
into one relation
 section(course_id, sec_id, semester, year,
building, room_number)
 No repetition in this case

D. NAGA JYOTHI CSE DEPT.

5

What About Smaller
Suppose we had started with inst_dept. How would we know to
split up Schemas?
(decompose) it into instructor and department?
 Write a rule “if there were a schema (dept_name, building,
budget), then dept_name would be a candidate key”
 Denote as a functional dependency:
dept_name → building, budget
 In inst_dept, because dept_name is not a candidate key, the
building and budget of a department may have to be repeated.
 This indicates the need to decompose inst_dept
 Not all decompositions are good. Suppose we decompose
employee(ID, name, street, city, salary) into
employee1 (ID, name)
employee2 (name, street, city, salary)
 The next slide shows how we lose information -- we cannot
reconstruct the original employee relation -- and so, this is a lossy
decomposition.
D. NAGA JYOTHI CSE DEPT.
6 A Lossy Decomposition

D. NAGA JYOTHI CSE DEPT.

Example of Lossless-Join Decomposition
7
 Lossless join decomposition
 Decomposition of R = (A, B, C)
R1 = (A, B) R2 = (B, C)

A B C A B B C
 1 A  1 1 A
 2 B  2 2 B
r A,B(r) B,C(r)

A B C
A (r) B (r)
 1 A
 2 B

D. NAGA JYOTHI CSE DEPT.

8 First Normal Form
 Domain is atomic if its elements are considered to be indivisible
units
 Examples of non-atomic domains:
 Set of names, composite attributes
 Identification numbers like CS101 that can be broken up into parts

 A relational schema R is in first normal form if the domains of all

attributes of R are atomic
 Non-atomic values complicate storage and encourage
redundant (repeated) storage of data
 Example: Set of accounts stored with each customer, and set of
owners stored with each account
 We assume all relations are in first normal form (and revisit this in
Chapter 22: Object Based Databases)

D. NAGA JYOTHI CSE DEPT.

First Normal Form (Cont’d)
9
 Atomicity is actually a property of how the elements of the
domain are used.
 Example: Strings would normally be considered indivisible
 Suppose that students are given roll numbers which are strings of the
form CS0012 or EE1127
 If the first two characters are extracted to find the department, the
domain of roll numbers is not atomic.
 Doing so is a bad idea: leads to encoding of information in
application program rather than in the database.

D. NAGA JYOTHI CSE DEPT.

Goal — Devise a Theory for the Following
10
 Decide whether a particular relation R is in “good” form.
 In the case that a relation R is not in “good” form, decompose it
into a set of relations {R1, R2, ..., Rn} such that
 each relation is in good form
 the decomposition is a lossless-join decomposition
 Our theory is based on:
 functional dependencies
 multivalued dependencies

D. NAGA JYOTHI CSE DEPT.

Functional Dependencies
11

 Constraints on the set of legal relations.

 Require that the value for a certain set of attributes
determines uniquely the value for another set of
attributes.
 A functional dependency is a generalization of the
notion of a key.

D. NAGA JYOTHI CSE DEPT.

12 Functional Dependencies
 Let R be a relation schema
(Cont.)
  R and   R
 The functional dependency
→
holds on R if and only if for any legal relations r(R), whenever
any two tuples t1 and t2 of r agree on the attributes , they
also agree on the attributes . That is,
t1[] = t2 []  t1[ ] = t2 [ ]
 Example: Consider r(A,B ) with the following instance of r.

1 4
1 5
3 7
 On this instance, A → B does NOT hold, but B → A does
hold.

D. NAGA JYOTHI CSE DEPT.

13 Functional Dependencies
 K is a superkey for relation schema R if and only if K → R
 (Cont.)
K is a candidate key for R if and only if
 K → R, and
 for no   K,  → R
 Functional dependencies allow us to express constraints that
cannot be expressed using superkeys. Consider the schema:
inst_dept (ID, name, salary, dept_name, building, budget ).
We expect these functional dependencies to hold:
dept_name→ building
and ID → building
but would not expect the following to hold:
dept_name → salary

D. NAGA JYOTHI CSE DEPT.

14 Use of Functional
We use functional dependencies to:
Dependencies

 test relations to see if they are legal under a given set of functional
dependencies.
 If a relation r is legal under a set F of functional dependencies, we say that r
satisfies F.

 specify constraints on the set of legal relations

 We say that F holds on R if all legal relations on R satisfy the set of functional
dependencies F.

 Note: A specific instance of a relation schema may satisfy a

functional dependency even if the functional dependency does
not hold on all legal instances.
 For example, a specific instance of instructor may, by chance, satisfy
name → ID.

D. NAGA JYOTHI CSE DEPT.

15 Functional Dependencies
(Cont.)
 A functional dependency is trivial if it is satisfied by all
instances of a relation
 Example:
 ID, name → ID
 name → name

 In general,  →  is trivial if   

D. NAGA JYOTHI CSE DEPT.

Closure
16 of a Set of Functional
Dependencies
 Given a set F of functional dependencies, there are certain
other functional dependencies that are logically implied by F.
 For example: If A → B and B → C, then we can infer that A → C
 The set of all functional dependencies logically implied by F is
the closure of F.
 We denote the closure of F by F+.
 F+ is a superset of F.

D. NAGA JYOTHI CSE DEPT.

17 Boyce-Codd Normal Form
A relation schema R is in BCNF with respect to a set F of
functional dependencies if for all functional dependencies
in F+ of the form

 →

where   R and   R, at least one of the following holds:

  →  is trivial (i.e.,   )
  is a superkey for R

Example schema not in BCNF:

instr_dept (ID, name, salary, dept_name, building, budget )

because dept_name→ building, budget

holds on instr_dept, but dept_name is not a superkey

D. NAGA JYOTHI CSE DEPT.

18 Decomposing a Schema
 Suppose we have a schema R and a non-trivial dependency  →
causesinto BCNF
a violation of BCNF.
We decompose R into:
• ( U  )
• (R-(-))
 In our example,
  = dept_name
  = building, budget
and inst_dept is replaced by
 ( U  ) = ( dept_name, building, budget )
 ( R - (  -  ) ) = ( ID, name, salary, dept_name )

D. NAGA JYOTHI CSE DEPT.

19 BCNF and Dependency
Preservation
 Constraints, including functional dependencies, are
costly to check in practice unless they pertain to only
one relation
 If it is sufficient to test only those dependencies on
each individual relation of a decomposition in order to
ensure that all functional dependencies hold, then
that decomposition is dependency preserving.
 Because it is not always possible to achieve both BCNF
and dependency preservation, we consider a weaker
normal form, known as third normal form.

D. NAGA JYOTHI CSE DEPT.

20 Third Normal Form
 A relation schema R is in third normal form (3NF) if for all:
 →  in F+
at least one of the following holds:
  →  is trivial (i.e.,   )
  is a superkey for R
 Each attribute A in  –  is contained in a candidate key for R.
(NOTE: each attribute may be in a different candidate key)
 If a relation is in BCNF it is in 3NF (since in BCNF one of the first two
conditions above must hold).
 Third condition is a minimal relaxation of BCNF to ensure
dependency preservation (will see why later).

D. NAGA JYOTHI CSE DEPT.

 Goals of Normalization
21
 Let R be a relation scheme with a set F of functional
dependencies.
 Decide whether a relation scheme R is in “good” form.
 In the case that a relation scheme R is not in “good” form,
decompose it into a set of relation scheme {R1, R2, ..., Rn} such
that
 each relation scheme is in good form
 the decomposition is a lossless-join decomposition
 Preferably, the decomposition should be dependency preserving.

D. NAGA JYOTHI CSE DEPT.

 How good is BCNF?
22

 There are database schemas in BCNF that do not

seem to be sufficiently normalized
 Consider a relation
inst_info (ID, child_name, phone)
ID child_name phone
 where an instructor may have more than one phone and
512-555-1234
99999 can have multiple children
David
512-555-4321
99999 David
512-555-1234
99999 William
512-555-4321
99999 Willian

inst_info
D. NAGA JYOTHI CSE DEPT.
 How good is BCNF? (Cont.)
23

 There are no non-trivial functional dependencies and therefore

the relation is in BCNF
 Insertion anomalies – i.e., if we add a phone 981-992-3443 to
99999, we need to add two tuples
(99999, David, 981-992-3443)
(99999, William, 981-992-3443)

D. NAGA JYOTHI CSE DEPT.

 How good is BCNF? (Cont.)
24
 Therefore, it is better to decompose inst_info into:

ID child_name

inst_child 99999 David

99999 David
99999 William
99999 Willian

ID phone
512-555-1234
inst_phone 99999
512-555-4321
99999
512-555-1234
99999
512-555-4321
99999

This suggests the need for higher normal forms, such as

Fourth Normal Form
D. NAGA JYOTHI (4NF), which we shall see later.
CSE DEPT.
25 Functional-Dependency
Theory
 We now consider the formal theory that tells us which
functional dependencies are implied logically by a
given set of functional dependencies.
 We then develop algorithms to generate lossless
decompositions into BCNF and 3NF
 We then develop algorithms to test if a decomposition
is dependency-preserving

D. NAGA JYOTHI CSE DEPT.

Closure
26 of a Set of Functional
Dependencies
 Given a set F set of functional dependencies, there are
certain other functional dependencies that are logically
implied by F.
 For e.g.: If A → B and B → C, then we can infer that A → C
 The set of all functional dependencies logically implied by F is
the closure of F.
 We denote the closure of F by F+.

D. NAGA JYOTHI CSE DEPT.

Closure
27 of a Set of Functional
Dependencies
 We can find F+, the closure of F, by repeatedly applying
Armstrong’s Axioms:
 if   , then  →  (reflexivity)
 if  →  , then   →   (augmentation)
 if  →  , and  → , then  →  (transitivity)
 These rules are
 sound (generate only functional dependencies that actually hold),
and
 complete (generate all functional dependencies that hold).

D. NAGA JYOTHI CSE DEPT.

 Example
28 R = (A, B, C, G, H, I)
F={ A→B
A→C
CG → H
CG → I
B → H}
 some members of F+
 A→H
 by transitivity from A → B and B → H

 AG → I
 by augmenting A → C with G, to get AG → CG
and then transitivity with CG → I

 CG → HI
 by augmenting CG → I to infer CG → CGI,
and augmenting of CG → H to infer CGI → HI,
and then transitivity

D. NAGA JYOTHI CSE DEPT.

29 Procedure for Computing F+
 To compute the closure of a set of functional dependencies F:

F+=F
repeat
for each functional dependency f in F+
apply reflexivity and augmentation rules on f
add the resulting functional dependencies to F +
for each pair of functional dependencies f1and f2 in F +
if f1 and f2 can be combined using transitivity
then add the resulting functional dependency to F +
until F + does not change any further

NOTE: We shall see an alternative procedure for this task later

D. NAGA JYOTHI CSE DEPT.

 Closure of Functional Dependencies
30
(Cont.)
 Additional rules:
 If  →  holds and  →  holds, then  →   holds (union)
 If  →   holds, then  →  holds and  →  holds
(decomposition)
 If  →  holds and   →  holds, then   →  holds
(pseudotransitivity)
The above rules can be inferred from Armstrong’s axioms.

D. NAGA JYOTHI CSE DEPT.

31 Closure of Attribute Sets

 Given a set of attributes , define the closure of 

under F (denoted by +) as the set of attributes that
are functionally determined by  under F

 Algorithm to compute +, the closure of  under F

result := ;
while (changes to result) do
for each  →  in F do
begin
if   result then result := result  
end

D. NAGA JYOTHI CSE DEPT.

 Example of Attribute Set
32  R = (A, B, C, G, H, I)

A Closure
 F = {A → B
→C
CG → H
CG → I
B → H}
 (AG)+
1. result = AG
2. result = ABCG (A → C and A → B)
3. result = ABCGH (CG → H and CG  AGBC)
4. result = ABCGHI (CG → I and CG  AGBCH)
 Is AG a candidate key?
1. Is AG a super key?
1. Does AG → R? == Is (AG)+  R
2. Is any subset of AG a superkey?
1. Does A → R? == Is (A)+  R
2. D. Does G →CSE
NAGA JYOTHI
R?DEPT.
== Is (G)+  R
 Uses of Attribute Closure
33
There are several uses of the attribute closure algorithm:
 Testing for superkey:
 To test if  is a superkey, we compute +, and check if + contains all
attributes of R.
 Testing functional dependencies
 To check if a functional dependency  →  holds (or, in other words,
is in F+), just check if   +.
 That is, we compute + by using attribute closure, and then check if
it contains .
 Is a simple and cheap test, and very useful
 Computing closure of F
 For each   R, we find the closure +, and for each S  +, we output
a functional dependency  → S.

D. NAGA JYOTHI CSE DEPT.

34 Canonical Cover

 Sets of functional dependencies may have redundant

dependencies that can be inferred from the others
 For example: A → C is redundant in: {A → B, B → C, A→
C}
 Parts of a functional dependency may be redundant
 E.g.: on RHS: {A → B, B → C, A → CD} can be simplified to
{A → B, B → C, A → D}
 E.g.: on LHS: {A → B, B → C, AC → D} can be simplified to
{A → B, B → C, A → D}

 Intuitively, a canonical cover of F is a “minimal” set of

functional dependencies equivalent to F, having no
redundant dependencies or redundant parts of
dependencies

D. NAGA JYOTHI CSE DEPT.

35 Extraneous Attributes
 Consider a set F of functional dependencies and the functional
dependency  →  in F.
 Attribute A is extraneous in  if A  
and F logically implies (F – { → })  {( – A) → }.
 Attribute A is extraneous in  if A  
and the set of functional dependencies
(F – { → })  { →( – A)} logically implies F.
 Note: implication in the opposite direction is trivial in each of
the cases above, since a “stronger” functional dependency
always implies a weaker one
 Example: Given F = {A → C, AB → C }
 B is extraneous in AB → C because {A → C, AB → C} logically implies
A → C (I.e. the result of dropping B from AB → C).
 Example: Given F = {A → C, AB → CD}
 C is extraneous in AB → CD since AB → C can be inferred even
after deleting C
D. NAGA JYOTHI CSE DEPT.
 Testing if an Attribute is Extraneous
36

 Consider a set F of functional dependencies and the

functional dependency  →  in F.
 To test if attribute A   is extraneous in 
1. compute ({} – A)+ using the dependencies in F
2. check that ({} – A)+ contains ; if it does, A is
extraneous in 
 To test if attribute A   is extraneous in 
1. compute + using only the dependencies in
F’ = (F – { → })  { →( – A)},
2. check that + contains A; if it does, A is extraneous in 

D. NAGA JYOTHI CSE DEPT.

37 Canonical Cover
 A canonical cover for F is a set of dependencies Fc such that
 F logically implies all dependencies in Fc, and
 Fc logically implies all dependencies in F, and
 No functional dependency in Fc contains an extraneous attribute, and
 Each left side of functional dependency in Fc is unique.
 To compute a canonical cover for F:
repeat
Use the union rule to replace any dependencies in F
1 → 1 and 1 → 2 with 1 → 1 2
Find a functional dependency  →  with an
extraneous attribute either in  or in 
/* Note: test for extraneous attributes done using Fc, not
F*/
If an extraneous attribute is found, delete it from  → 
until F does not change
 Note: Union rule may become applicable after some extraneous
attributes have been deleted, so it has to be re-applied
D. NAGA JYOTHI CSE DEPT.
 Computing a Canonical Cover
38  R = (A, B, C)
F = {A → BC
B→C
A→B
AB → C}
 Combine A → BC and A → B into A → BC
 Set is now {A → BC, B → C, AB → C}
 A is extraneous in AB → C
 Check if the result of deleting A from AB → C is implied by the other
dependencies
 Yes: in fact, B → C is already present!
 Set is now {A → BC, B → C}
 C is extraneous in A → BC
 Check if A → C is logically implied by A → B and the other dependencies
 Yes: using transitivity on A → B and B → C.
 Can use attribute closure of A in more complex cases
 The canonical cover is: A→B
B→C
D. NAGA JYOTHI CSE DEPT.
39 Lossless-join Decomposition
 For the case of R = (R1, R2), we require that for all possible
relations r on schema R
r = R1 (r ) R2 (r )
 A decomposition of R into R1 and R2 is lossless join if at least one
of the following dependencies is in F+:
 R1  R2 → R1
 R1  R2 → R2
 The above functional dependencies are a sufficient condition
for lossless join decomposition; the dependencies are a
necessary condition only if all constraints are functional
dependencies

D. NAGA JYOTHI CSE DEPT.

40 Example
 R = (A, B, C)
F = {A → B, B → C)
 Can be decomposed in two different ways
 R1 = (A, B), R2 = (B, C)
 Lossless-join decomposition:
R1  R2 = {B} and B → BC
 Dependency preserving
 R1 = (A, B), R2 = (A, C)
 Lossless-join decomposition:
R1  R2 = {A} and A → AB
 Not dependency preserving
(cannot check B → C without computing R1 R2)

D. NAGA JYOTHI CSE DEPT.

Dependency Preservation
41
 Let Fi be the set of dependencies F + that include only
attributes in Ri.
 A decomposition is dependency preserving, if
(F  F  …  F )+ = F +
1 2 n

 If it is not, then checking updates for violation of functional

dependencies may require computing joins, which is expensive.

D. NAGA JYOTHI CSE DEPT.

Testing for Dependency Preservation
42
 To check if a dependency  →  is preserved in a
decomposition of R into R1, R2, …, Rn we apply the following
test (with attribute closure done with respect to F)
 result = 
while (changes to result) do
for each Ri in the decomposition
t = (result  Ri)+  Ri
result = result  t
 If result contains all attributes in , then the functional
dependency
 →  is preserved.
 We apply the test on all dependencies in F to check if a
decomposition is dependency preserving
 This procedure takes polynomial time, instead of the
exponential time required to compute F+ and (F1  F2  … 
Fn)+

D. NAGA JYOTHI CSE DEPT.

43 Example
 R = (A, B, C )
F = {A → B
B → C}
Key = {A}
 R is not in BCNF
 Decomposition R1 = (A, B), R2 = (B, C)
 R1 and R2 in BCNF
 Lossless-join decomposition
 Dependency preserving

D. NAGA JYOTHI CSE DEPT.

44

Testing for BCNF
To check if a non-trivial dependency  → causes a violation of
BCNF
1. compute + (the attribute closure of ), and
2. verify that it includes all attributes of R, that is, it is a superkey of R.
 Simplified test: To check if a relation schema R is in BCNF, it
suffices to check only the dependencies in the given set F for
violation of BCNF, rather than checking all dependencies in F+.
 If none of the dependencies in F causes a violation of BCNF, then
none of the dependencies in F+ will cause a violation of BCNF either.
 However, simplified test using only F is incorrect when testing a
relation in a decomposition of R
 Consider R = (A, B, C, D, E), with F = { A → B, BC → D}
 Decompose R into R1 = (A,B) and R2 = (A,C,D, E)
 Neither of the dependencies in F contain only attributes from
(A,C,D,E) so we might be mislead into thinking R2 satisfies BCNF.
 In fact, dependency AC → D in F+ shows R2 is not in BCNF.

D. NAGA JYOTHI CSE DEPT.

45 Testing Decomposition for
To check if a relation R in a decomposition of R is in BCNF,
BCNF
 i

 Either test R for BCNF with respect to the restriction of F to R (that is,
i i
all FDs in F+ that contain only attributes from Ri)
 or use the original set of dependencies F that hold on R, but with
the following test:
 for every set of attributes   Ri, check that + (the attribute closure of ) either
includes no attribute of Ri- , or includes all attributes of Ri.

 If the condition is violated by some  →  in F, the dependency

 → (+ -  )  Ri
can be shown to hold on Ri, and Ri violates BCNF.
 We use above dependency to decompose Ri

D. NAGA JYOTHI CSE DEPT.

46 BCNF Decomposition
Algorithm
result := {R };
done := false;
compute F +;
while (not done) do
if (there is a schema Ri in result that is not in BCNF)
then begin
let  →  be a nontrivial functional dependency that
holds on Ri such that  → Ri is not in F +,
and    = ;
result := (result – Ri )  (Ri – )  (,  );
end
else done := true;

Note: each Ri is in BCNF, and decomposition is lossless-join.

D. NAGA JYOTHI CSE DEPT.

47 Example of BCNF
 R = (A, B, C )
F = {A →Decomposition
B
B → C}
Key = {A}
 R is not in BCNF (B → C but B is not superkey)
 Decomposition
 R1 = (B, C)
 R2 = (A,B)

D. NAGA JYOTHI CSE DEPT.

48 Example of BCNF Decomposition
 class (course_id, title, dept_name, credits, sec_id, semester, year,
building, room_number, capacity, time_slot_id)
 Functional dependencies:
 course_id→ title, dept_name, credits
 building, room_number→capacity
 course_id, sec_id, semester, year→building, room_number,
time_slot_id
 A candidate key {course_id, sec_id, semester, year}.
 BCNF Decomposition:
 course_id→ title, dept_name, credits holds
 but course_id is not a superkey.
 We replace class by:
 course(course_id, title, dept_name, credits)
 class-1 (course_id, sec_id, semester, year, building,
room_number, capacity, time_slot_id)

D. NAGA JYOTHI CSE DEPT.

49 BCNF Decomposition (Cont.)

 course is in BCNF
 How do we know this?
 building, room_number→capacity holds on class-1
 but {building, room_number} is not a superkey for class-1.
 We replace class-1 by:
 classroom (building, room_number, capacity)
 section (course_id, sec_id, semester, year, building,
room_number, time_slot_id)

 classroom and section are in BCNF.

D. NAGA JYOTHI CSE DEPT.

 BCNF and Dependency Preservation
50
It is not always possible to get a BCNF decomposition that is
dependency preserving

 R = (J, K, L )
F = {JK → L
L→K}
Two candidate keys = JK and JL
 R is not in BCNF
 Any decomposition of R will fail to preserve
JK → L
This implies that testing for JK → L requires a join

D. NAGA JYOTHI CSE DEPT.

51 Third Normal Form:
 There are some situations where
 BCNFMotivation
is not dependency preserving, and
 efficient checking for FD violation on updates is important
 Solution: define a weaker normal form, called Third
Normal Form (3NF)
 Allows some redundancy (with resultant problems; we will see
examples later)
 But functional dependencies can be checked on individual
relations without computing a join.
 There is always a lossless-join, dependency-preserving
decomposition into 3NF.

D. NAGA JYOTHI CSE DEPT.

52 3NF Example
 Relation dept_advisor:
 dept_advisor (s_ID, i_ID, dept_name)
F = {s_ID, dept_name → i_ID, i_ID → dept_name}
 Two candidate keys: s_ID, dept_name, and i_ID, s_ID
 R is in 3NF
 s_ID, dept_name → i_ID s_ID
 dept_name is a superkey

 i_ID → dept_name
 dept_name is contained in a candidate key

D. NAGA JYOTHI CSE DEPT.

 Redundancy in 3NF
53
 There is some redundancy in this schema
 Example of problems due to redundancy in 3NF
 R = (J, K, L)
F = {JK → L, L → K } J L K
j1 l1 k1
j2 l1 k1
j3 l1 k1

null l2 k2

repetition of information (e.g., the relationship l1, k1)

⚫ (i_ID, dept_name)
need to use null values (e.g., to represent the relationship
l2, k2 where there is no corresponding value for J).
⚫ (i_ID, dept_nameI) if there is no separate relation mapping
instructors to departments
D. NAGA JYOTHI CSE DEPT.
54 Testing for 3NF

 Optimization: Need to check only FDs in F, need not

check all FDs in F+.
 Use attribute closure to check for each dependency 
→ , if  is a superkey.
 If  is not a superkey, we have to verify if each
attribute in  is contained in a candidate key of R
 this test is rather more expensive, since it involve finding
candidate keys
 testing for 3NF has been shown to be NP-hard
 Interestingly, decomposition into third normal form
(described shortly) can be done in polynomial time

D. NAGA JYOTHI CSE DEPT.

55 3NF Decomposition
Let F be a canonical cover for F;
i := 0; Algorithm
c

for each functional dependency  →  in Fc do

if none of the schemas Rj, 1  j  i contains  
then begin
i := i + 1;
Ri :=  
end
if none of the schemas Rj, 1  j  i contains a candidate key for R
then begin
i := i + 1;
Ri := any candidate key for R;
end
/* Optionally, remove redundant relations */
repeat
if any schema Rj is contained in another schema Rk
then /* delete Rj */
Rj = R;;
i=i-1;
return (R1, R2, ..., Ri)
D. NAGA JYOTHI CSE DEPT.
 3NF Decomposition Algorithm (Cont.)
56

 Above algorithm ensures:

 each relation schema Ri is in 3NF
 decomposition is dependency preserving and lossless-
join
 Proof of correctness is at end of this presentation (click
here)

D. NAGA JYOTHI CSE DEPT.

57 3NF Decomposition: An
Relation schema:
Example

cust_banker_branch = (customer_id, employee_id, branch_name, type )
 The functional dependencies for this relation schema are:
1. customer_id, employee_id → branch_name, type
2. employee_id → branch_name
3. customer_id, branch_name → employee_id
 We first compute a canonical cover
 branch_name is extraneous in the r.h.s. of the 1st dependency
 No other attribute is extraneous, so we get FC =
customer_id, employee_id → type
employee_id → branch_name
customer_id, branch_name → employee_id

D. NAGA JYOTHI CSE DEPT.

58 3NF Decompsition Example
 The for loop generates following 3NF schema:
(Cont.)employee_id, type )
(customer_id,
(employee_id, branch_name)
(customer_id, branch_name, employee_id)
 Observe that (customer_id, employee_id, type ) contains a candidate
key of the original schema, so no further relation schema needs be
added
 At end of for loop, detect and delete schemas, such as
(employee_id, branch_name), which are subsets of other
schemas
 result will not depend on the order in which FDs are considered

 The resultant simplified 3NF schema is:

(customer_id, employee_id, type)
(customer_id, branch_name, employee_id)

D. NAGA JYOTHI CSE DEPT.

59 Comparison of BCNF and
3NF
 It is always possible to decompose a relation into a set
of relations that are in 3NF such that:
 the decomposition is lossless
 the dependencies are preserved
 It is always possible to decompose a relation into a set
of relations that are in BCNF such that:
 the decomposition is lossless
 it may not be possible to preserve dependencies.

D. NAGA JYOTHI CSE DEPT.

60

Design Goals
Goal for a relational database design is:
 BCNF.
 Lossless join.
 Dependency preservation.
 If we cannot achieve this, we accept one of
 Lack of dependency preservation
 Redundancy due to use of 3NF
 Interestingly, SQL does not provide a direct way of specifying
functional dependencies other than superkeys.
Can specify FDs using assertions, but they are expensive to test,
(and currently not supported by any of the widely used databases!)
 Even if we had a dependency preserving decomposition, using SQL
we would not be able to efficiently test a functional dependency
whose left hand side is not a key.

D. NAGA JYOTHI CSE DEPT.

61 Multivalued Dependencies

 Suppose we record names of children, and phone

numbers for instructors:
 inst_child(ID, child_name)
 inst_phone(ID, phone_number)
 If we were to combine these schemas to get
 inst_info(ID, child_name, phone_number)
 Example data:
(99999, David, 512-555-1234)
(99999, David, 512-555-4321)
(99999, William, 512-555-1234)
(99999, William, 512-555-4321)
 This relation is in BCNF
 Why?

D. NAGA JYOTHI CSE DEPT.

62 Multivalued Dependencies

(MVDs)
Let R be a relation schema and let   R and   R. The
multivalued dependency
 →→ 
holds on R if in any legal relation r(R), for all pairs for tuples t1
and t2 in r such that t1[] = t2 [], there exist tuples t3 and t4 in r
such that:
t1[] = t2 [] = t3 [] = t4 []
t3[] = t1 []
t3[R – ] = t2[R – ]
t4 [] = t2[]
t4[R – ] = t1[R – ]

D. NAGA JYOTHI CSE DEPT.

63 MVD (Cont.)
 Tabular representation of  →→ 

D. NAGA JYOTHI CSE DEPT.

64 Example

 Let R be a relation schema with a set of attributes that

are partitioned into 3 nonempty subsets.
Y, Z, W
 We say that Y →→ Z (Y multidetermines Z )
if and only if for all possible relations r (R )
< y1, z1, w1 >  r and < y1, z2, w2 >  r
then
< y1, z1, w2 >  r and < y1, z2, w1 >  r
 Note that since the behavior of Z and W are identical
it follows that
Y →→ Z if Y →→ W

D. NAGA JYOTHI CSE DEPT.

65 Example (Cont.)
 In our example:
ID →→ child_name
ID →→ phone_number
 The above formal definition is supposed to formalize the notion
that given a particular value of Y (ID) it has associated with it a set
of values of Z (child_name) and a set of values of W
(phone_number), and these two sets are in some sense
independent of each other.
 Note:
 If Y → Z then Y →→ Z
 Indeed we have (in above notation) Z1 = Z2
The claim follows.

D. NAGA JYOTHI CSE DEPT.

66 Use of Multivalued
 We use multivalued dependencies in two ways:
Dependencies
1. To test relations to determine whether they are legal under a
given set of functional and multivalued dependencies
2. To specify constraints on the set of legal relations. We shall thus
concern ourselves only with relations that satisfy a given set of
functional and multivalued dependencies.
 If a relation r fails to satisfy a given multivalued dependency,
we can construct a relations r that does satisfy the
multivalued dependency by adding tuples to r.

D. NAGA JYOTHI CSE DEPT.

67 Theory of MVDs
 From the definition of multivalued dependency, we can derive
the following rule:
 If  → , then  →→ 
That is, every functional dependency is also a multivalued
dependency
 The closure D+ of D is the set of all functional and multivalued
dependencies logically implied by D.
 We can compute D+ from D, using the formal definitions of functional
dependencies and multivalued dependencies.
 We can manage with such reasoning for very simple multivalued
dependencies, which seem to be most common in practice
 For complex dependencies, it is better to reason about sets of
dependencies using a system of inference rules (see Appendix C).

D. NAGA JYOTHI CSE DEPT.

68 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  →→ , where   R and   R,
at least one of the following hold:
  →→  is trivial (i.e.,    or    = R)
  is a superkey for schema R
 If a relation is in 4NF it is in BCNF

D. NAGA JYOTHI CSE DEPT.

 Restriction of Multivalued
69Dependencies

 The restriction of D to Ri is the set Di consisting of

 All functional dependencies in D+ that include only
attributes of Ri
 All multivalued dependencies of the form
 →→ (  Ri)

where   Ri and  →→  is in D+

D. NAGA JYOTHI CSE DEPT.

70 4NF Decomposition
done :=Algorithm
result: = {R};
false;
compute D+;
Let Di denote the restriction of D+ to Ri
while (not done)
if (there is a schema Ri in result that is not in 4NF) then
begin
let  →→  be a nontrivial multivalued dependency that holds
on Ri such that  → Ri is not in Di, and =;
result := (result - Ri)  (Ri - )  (, );
end
else done:= true;
Note: each Ri is in 4NF, and decomposition is lossless-join

D. NAGA JYOTHI CSE DEPT.

 71 Example
 R =(A, B, C, G, H, I)
F ={ A →→ B
B →→ HI
CG →→ H }
 R is not in 4NF since A →→ B and A is not a superkey for R
 Decomposition
a) R1 = (A, B) (R1 is in 4NF)
b) R2 = (A, C, G, H, I) (R2 is not in 4NF, decompose into R3 and R4)
c) R3 = (C, G, H) (R3 is in 4NF)
d) R4 = (A, C, G, I) (R4 is not in 4NF, decompose into R5 and R6)
 A →→ B and B →→ HI ➔ A →→ HI, (MVD transitivity), and
 and hence A →→ I (MVD restriction to R4)
e) R5 = (A, I) (R5 is in 4NF)
f)R6 = (A, C, G) (R6 is in 4NF)

D. NAGA JYOTHI CSE DEPT.

Further Normal Forms
72
 Join dependencies generalize multivalued dependencies
 lead to project-join normal form (PJNF) (also called fifth normal
form)
 A class of even more general constraints, leads to a normal form
called domain-key normal form.
 Problem with these generalized constraints: are hard to reason
with, and no set of sound and complete set of inference rules
exists.
 Hence rarely used

D. NAGA JYOTHI CSE DEPT.

73 Overall Database Design
We have assumed schema R is given
Process

 R could have been generated when converting E-R diagram to a set of
tables.
 R could have been a single relation containing all attributes that are of
interest (called universal relation).
 Normalization breaks R into smaller relations.
 R could have been the result of some ad hoc design of relations, which
we then test/convert to normal form.

D. NAGA JYOTHI CSE DEPT.

74 ER Model and Normalization
 When an E-R diagram is carefully designed, identifying all entities
correctly, the tables generated from the E-R diagram should not
need further normalization.
 However, in a real (imperfect) design, there can be functional
dependencies from non-key attributes of an entity to other
attributes of the entity
 Example: an employee entity with attributes
department_name and building,
and a functional dependency
department_name→ building
 Good design would have made department an entity
 Functional dependencies from non-key attributes of a relationship
set possible, but rare --- most relationships are binary

D. NAGA JYOTHI CSE DEPT.

75 Denormalization for Performance
 May want to use non-normalized schema for performance
 For example, displaying prereqs along with course_id, and title
requires join of course with prereq
 Alternative 1: Use denormalized relation containing attributes of
course as well as prereq with all above attributes
 faster lookup
 extra space and extra execution time for updates
 extra coding work for programmer and possibility of error in extra code
 Alternative 2: use a materialized view defined as
course prereq
 Benefits and drawbacks same as above, except no extra coding work
for programmer and avoids possible errors

D. NAGA JYOTHI CSE DEPT.

76 Other Design Issues
 Some aspects of database design are not caught by normalization
 Examples of bad database design, to be avoided:
Instead of earnings (company_id, year, amount ), use
 earnings_2004, earnings_2005, earnings_2006, etc., all on the schema
(company_id, earnings).
 Above are in BCNF, but make querying across years difficult and needs new
table each year

 company_year (company_id, earnings_2004, earnings_2005,

earnings_2006)
 Also in BCNF, but also makes querying across years difficult and requires new
attribute each year.
 Is an example of a crosstab, where values for one attribute become column
names
 Used in spreadsheets, and in data analysis tools

D. NAGA JYOTHI CSE DEPT.

 Proof of Correctness
of 3NF
Decomposition
77
Algorithm

D. NAGA JYOTHI CSE DEPT.

Correctness of 3NF Decomposition
Algorithm
78

 3NF decomposition algorithm is dependency preserving (since

there is a relation for every FD in Fc)
 Decomposition is lossless
 A candidate key (C ) is in one of the relations Ri in decomposition
 Closure of candidate key under Fc must contain all attributes in R.
 Follow the steps of attribute closure algorithm to show there is only
one tuple in the join result for each tuple in Ri

D. NAGA JYOTHI CSE DEPT.

Correctness of 3NF Decomposition
Algorithm
79
(Cont’d.)

Claim: if a relation Ri is in the decomposition generated by the

above algorithm, then Ri satisfies 3NF.
 Let Ri be generated from the dependency  → 
 Let  → B be any non-trivial functional dependency on Ri. (We
need only consider FDs whose right-hand side is a single attribute.)
 Now, B can be in either  or  but not in both. Consider each case
separately.

D. NAGA JYOTHI CSE DEPT.

Correctness of 3NF Decomposition
(Cont’d.)
80
 Case 1: If B in :
 If  is a superkey, the 2nd condition of 3NF is satisfied
 Otherwise  must contain some attribute not in 
 Since  → B is in F+ it must be derivable from Fc, by using attribute
closure on .
 Attribute closure not have used  →. If it had been used,  must be
contained in the attribute closure of , which is not possible, since we
assumed  is not a superkey.
 Now, using → (- {B}) and  → B, we can derive  →B
(since    , and B   since  → B is non-trivial)
 Then, B is extraneous in the right-hand side of  →; which is not
possible since  → is in Fc.
 Thus, if B is in  then  must be a superkey, and the second condition
of 3NF must be satisfied.

D. NAGA JYOTHI CSE DEPT.

Correctness of 3NF Decomposition
(Cont’d.)
81
 Case 2: B is in .
 Since  is a candidate key, the third alternative in the definition of 3NF
is trivially satisfied.
 In fact, we cannot show that  is a superkey.
 This shows exactly why the third alternative is present in the definition
of 3NF.
Q.E.D.

D. NAGA JYOTHI CSE DEPT.

82 Figure 8.02

D. NAGA JYOTHI CSE DEPT.

83 Figure 8.03

D. NAGA JYOTHI CSE DEPT.

84 Figure 8.04

D. NAGA JYOTHI CSE DEPT.

85 Figure 8.05

D. NAGA JYOTHI CSE DEPT.

86 Figure 8.06

D. NAGA JYOTHI CSE DEPT.

87 Figure 8.14

D. NAGA JYOTHI CSE DEPT.

88 Figure 8.15

D. NAGA JYOTHI CSE DEPT.

89 Figure 8.17

D. NAGA JYOTHI CSE DEPT.