RELATIONAL DATABASE DESIGN
Relational Database Design
First Normal Form
Pitfalls in Relational Database Design
Functional Dependencies
Decomposition
Boyce-Codd Normal Form
Third Normal Form
Multivalued Dependencies and Fourth Normal
Form
Overall Database Design Process
Pitfalls in Relational Database Design
Relational database design requires that we
find a “good” collection of relation schemas.
A bad design may lead to
Repetition of Information.
Inability to represent certain information.
Design Goals:
Avoid redundant data
Ensure that relationships among attributes are
represented
Facilitate the checking of updates for violation of
database integrity constraints.
Example
Consider the relation schema:
Lending-schema = (branch-name, branch-city, assets,
customer-name, loan-number, amount)
Redundancy:
Data for branch-name, branch-city, assets are repeated for each loan that a branch
makes
Wastes space
Complicates updating, introducing possibility of inconsistency of assets value
Decomposition
Null values
Cannot store information about a branch if no loans exist
Can use null values, but they are difficult to handle.
Decompose the relation schema Lending-schema into:
Branch-schema = (branch-name, branch-city,assets)
Loan-info-schema = (customer-name, loan-number,
branch-name,
amount)
All attributes of an original schema (R) must appear in
the decomposition (R1, R2):
R = R1 R2
Lossless-join decomposition.
For all possible relations r on schema R
r = R1 (r) R2 (r)
Example of Non Lossless-Join Decomposition
Decomposition of R = (A, B)
R1 = (A) R2 = (B)
A B A B
1 1
2 2
1 B(r)
A(r)
r
A B
A (r) B (r)
1
2
1
2
Goal — Devise a Theory for the
Following
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:
functionaldependencies
multivalued dependencies
Functional Dependencies
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.
Functional Dependencies (Cont.)
Let R be a relation schema
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.
Functional Dependencies (Cont.)
K is a superkey for relation schema R if and only if K R
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:
Loan-info-schema = (customer-name, loan-number,
branch-name, amount).
We expect this set of functional dependencies to hold:
loan-number amount
loan-number branch-name
but would not expect the following to hold:
loan-number customer-name
Use of Functional Dependencies
We use functional dependencies to:
testrelations 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.
For example, a specific instance of Loan-schema may,
by chance, satisfy
loan-number customer-name.
Functional Dependencies (Cont.)
A functional dependency is trivial if it is satisfied by
all instances of a relation
E.g.
customer-name, loan-number customer-name
customer-name customer-name
In general, is trivial if
Closure 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.
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+.
Arm Strong's Inference Rules/
axioms
if , then (reflexivity)
if , then (augmentation)
if , and , then (transitivity)
If CG H and CG I => CG HI
By “union rule”
If CG->H then C->H and G->H by
decomposition rule.
We can find all of F+ by 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).
Example
R = (A, B, C, G, H, I)
F={ AB
AC
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
from CG H and CG I : “union rule” can
be inferred from
definition
of functional dependencies, or
Augmentation of CG I to infer CG CGI,
augmentation of
CG H to infer CGI HI, and then transitivity
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+
NOTE: We will see an alternative procedure
for this task later
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
Closure of Functional Dependencies
(Cont.)
We can further simplify manual computation
of F+ by using the following 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.
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:
is in 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
Example of Attribute Set Closure
R = (A, B, C, G, H, I)
F = {A B
AC
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. Does G R? == Is (G)+ R
Uses of Attribute Closure
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
Foreach R, we find the closure +, and for each S
+, we output a functional dependency S.
Canonical Cover
Sets of functional dependencies may have
redundant dependencies that can be inferred from
the others
Eg: 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
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
Testing if an Attribute is Extraneous
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 A; if it does, A is
extraneous
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
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
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
Example of Computing a Canonical
Cover
R = (A, B, C)
F = {A BC
BC
AB
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: AB
BC
Goals of Normalization
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:
functionaldependencies
multivalued dependencies
Decomposition
Decompose the relation schema Lending-schema into:
Branch-schema = (branch-name, branch-city,assets)
Loan-info-schema = (customer-name, loan-number,
branch-name, amount)
All attributes of an original schema (R) must appear in the decomposition
(R1, R2):
R = R 1 R2
Lossless-join decomposition.
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 and only if at least
one of the following dependencies is in F+:
R1 R2 R1
R1 R2 R2
Example of Lossy-Join Decomposition
Lossy-join decompositions result in
information loss.
Example: Decomposition of R = (A, B)
A B R1 =A (A) R2 = (B)B
1 1
2 2
1 B(r)
A(r)
r
A B
A (r) B (r)
1
2
1
2
Normalization Using Functional Dependencies
When we decompose a relation schema R with a set of
functional dependencies F into R1, R2,.., Rn we want
Lossless-joindecomposition: Otherwise decomposition
would result in information loss.
No redundancy: The relations Ri preferably should be
in either Boyce-Codd Normal Form or Third Normal
Form.
Dependency preservation: Let Fi be the set
of dependencies F+ that include only
attributes in Ri.
Preferably the decomposition should be
dependency preserving, that is, (F1
F2 … Fn)+ = F+
Otherwise, checking updates for violation
of functional dependencies may require
computing joins, which is expensive.
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 R R)
Testing for Dependency
Preservation
To check if a dependency is preserved in a
decomposition of R into R1, R2, …, Rn we apply the
following simplified test (with attribute closure done
w.r.t. 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
F 2 … F n )+
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
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
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+.
Ifnone 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, using only F is incorrect when testing a
relation in a decomposition of R
E.g. Consider R (A, B, C, D), with F = { A B, B C}
Decompose R into R1(A,B) and R2(A,C,D)
Neither of the dependencies in F contain only attributes from
(A,C,D) so we might be mislead into thinking R2 satisfies
BCNF.
In fact, dependency A C in F+ shows R2 is not in BCNF.
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.
Example of BCNF Decomposition
R = (branch-name, branch-city, assets,
customer-name, loan-number, amount)
F = {branch-name assets branch-city
loan-number amount branch-name}
Key = {loan-number, customer-name}
Decomposition
R1 = (branch-name, branch-city, assets)
R2 = (branch-name, customer-name, loan-number,
amount)
R3 = (branch-name, loan-number, amount)
R4 = (customer-name, loan-number)
Final decomposition
Testing Decomposition for BCNF
To check if a relation Ri in a decomposition of R is
in BCNF,
Either test Ri for BCNF with respect to the restriction of
F to Ri (that is, all FDs in F+ that contain only attributes
from Ri)
BCNF and Dependency
Preservation
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
Third Normal Form: Motivation
There are some situations where
BCNF is not dependency preserving, and
efficient checking for FD violation on updates is
important
Solution: define a weaker normal form, called Third
Normal Form.
Allows some redundancy (with resultant problems; we
will see examples later)
But FDs can be checked on individual relations without
computing a join.
There is always a lossless-join, dependency-preserving
decomposition into 3NF.
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).
3NF (Cont.)
Example
R = (J, K, L)
F = {JK L, L K}
Two candidate keys: JK and JL
R is in 3NF
JK L JK is a superkey
LK K is contained in a candidate key
BCNF decomposition has (JL) and (LK)
Testing for JK L requires a join
There is some redundancy in this schema
Equivalent to example in book:
Banker-schema = (branch-name, customer-name, banker-name)
banker-name branch name
branch name customer-name banker-name
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
3NF Decomposition Algorithm
Let Fc be a canonical cover for F;
i := 0;
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;
3NF Decomposition Algorithm
(Cont.)
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 file (click here)
Example
Relation schema:
Banker-info-schema = (branch-name, customer-name,
banker-name, office-number)
The functional dependencies for this relation
schema are:
banker-name branch-name office-number
customer-name branch-name banker-name
The key is:
{customer-name, branch-name}
Applying 3NF to Banker-info-
schema
The for loop in the algorithm causes us to
include the following schemas in our
decomposition:
Banker-office-schema = (banker-name,
branch-name,
office-number)
Banker-schema = (customer-name, branch-
name,
banker-name)
Since Banker-schema contains a candidate key
for
Comparison of BCNF and 3NF
It is always possible to decompose a relation into
relations in 3NF and
the decomposition is lossless
the dependencies are preserved
It is always possible to decompose a relation into
relations in BCNF and
the decomposition is lossless
it may not be possible to preserve dependencies.
Comparison of BCNF and 3NF
(Cont.)
Example of problems due to redundancy in 3NF
R = (J, K, L)J L K
F = {JK jL,1 L l
1 kK}
1
j2 l1 k1
j3 l1 k1
null l2 k2
A schema that is in 3NF but not in BCNF has the problems of
repetition of information (e.g., the relationship l1, k1)
need to use null values (e.g., to represent the relationship
l2, k2 where there is no corresponding value for J).
Design Goals
Goal for a relational database design is:
BCNF.
Lossless
join.
Dependency preservation.
If we cannot achieve this, we accept one of
Lackof 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
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.
Testing for FDs Across Relations
If decomposition is not dependency preserving, we can have an extra
materialized view for each dependency in Fc that is not preserved in the
decomposition
The materialized view is defined as a projection on of the join of the relations
in the decomposition
Many newer database systems support materialized views and database system
maintains the view when the relations are updated.
No extra coding effort for programmer.
The functional dependency is expressed by declaring as a candidate
key on the materialized view.
Checking for candidate key cheaper than checking
BUT:
Space overhead: for storing the materialized view
Time overhead: Need to keep materialized view up to date when
relations are updated
Database system may not support key declarations on
materialized views
Multivalued Dependencies
There are database schemas in BCNF that do not
seem to be sufficiently normalized
Consider a database
classes(course, teacher, book)
such that (c,t,b) classes means that t is qualified to
teach c, and b is a required textbook for c
The database is supposed to list for each course the
set of teachers any one of which can be the course’s
instructor, and the set of books, all of which are
required for the course (no matter who teaches it).
Multivalued Dependencies
(Cont.)
course teacher book
database Avi DB Concepts
database Avi Ullman
database Hank DB Concepts
database Hank Ullman
database Sudarshan DB Concepts
database Sudarshan Ullman
operating systems Avi OS Concepts
operating systems Avi Shaw
operating systems Jim OS Concepts
operating systems Jim Shaw
classes
There are no non-trivial functional
dependencies and therefore the relation
is in BCNF
Insertion anomalies – i.e., if Sara is a
new teacher that can teach database,
two tuples need to be inserted
(database, Sara, DB Concepts)
(database, Sara, Ullman)
Multivalued Dependencies
(Cont.)
Therefore, it is better to decompose
classes into:
course teacher
database Avi
database Hank
database Sudarshan
operating systems Avi
operating systems Jim
teaches
course book
database DB Concepts
database Ullman
operating systems OS Concepts
operating systems Shaw
text
We shall see that these two relations are in Fourth Normal Form (4NF)
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 – ]
MVD (Cont.)
Tabular representation of
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 < y2, z2, w2 > r
then
< y1, z1, w2 > r and < y2, z2, w1 > r
Note that since the behavior of Z and W are
identical it follows that Y Z if Y W
Example (Cont.)
In our example:
course teacher
course book
The above formal definition is supposed to
formalize the notion that given a particular
value of Y (course) it has associated with it
a set of values of Z (teacher) and a set of
values of W (book), and these two sets are
in some sense independent of each other.
Use of Multivalued Dependencies
We use multivalued dependencies in two
ways:
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.
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).
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
Restriction of Multivalued
Dependencies
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+
4NF Decomposition Algorithm
result: = {R};
done := 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
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)
c) R3 = (C, G, H) (R3 is in 4NF)
d) R4 = (A, C, G, I) (R4 is not in 4NF)
Since A B and B HI, A HI, A I
e) R5 = (A, I) (R5 is in 4NF)
f)R6 = (A, C, G) (R6 is in 4NF)
Further Normal Forms
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.
