Chapter 7: Normalization

Question 1

What does decomposition do?

Answer 1

reduce repetition

Question 2

lossy decomposition

Answer 2

loss of information

Question 3

lossless Decomposition

Answer 3

includes all attributes of R

Question 4

Functional Dependencies

Answer 4

generalization of the notion of a key

Question 5

legal instance

Answer 5

relation satisfies all real world constraints

Question 6

B can hold A but…

Answer 6

sometimes A cannot hold B

Question 7

trivial

Answer 7

A gives B and B is a subset of A

Question 8

Trivial Examples:

Answer 8

ID, NAME gives ID
name gives name

Question 9

Closure

Answer 9

when all functional dependencies can be implied

Question 10

SuperKey

Answer 10

sufficient to identify a unique tuple

Question 11

Candidate Key

Answer 11

it is minimal, meaning just one key alone

Question 12

satisfies

Answer 12

relation r is legal under set F functional dependency

Question 13

holds on

Answer 13

all legal relations on R satisfies F functional dependency

Question 14

Visual of lossless Decomposition

Answer 14

R1 and R2 gives R1

R2 and R1 gives R2

Question 15

First Normal Form

Answer 15

Domain is atomic

Question 16

Atomic

Answer 16

indivisible units
ex: names, or ID numbers

Question 17

non-atomic numbers can..

Answer 17

• complicate storage
• cause redundancy

Question 18

When decomposing a relation, to do testing for dependency preservation you cannot use …

Answer 18

Cartesian product or joins

Question 19

Boyce - Codd Normal Form

Answer 19

at least One of the following holds true…

a to beta is trivial
a is a superkey of R

Question 20

Decomposing into BCNF

Answer 20

R1 = a or B
R2 = R - (B - a)

Question 21

Not always possible to achieve both BCNF and…

Answer 21

Dependency Preservation

Question 22

Third Normal Form is only if…

Answer 22

• a gives B is trivial
• a is a superkey for R
• attributes in B-a is in the candidate key

Question 23

Advantages of 3NF versus BCNF

Answer 23

3NF always has dependency preservation

Question 24

Disadvantages of 3NF

Answer 24

can have null values
redundancy

Question 25

Armstrong Axiom's

Answer 25

* Reflexive Rule is B is part of A then A gives B * Augmentation rule is a gives b then ya gives yb *Transitivity rule is a gives b and b gives y then a gives y

Question 26

These rules are...

Answer 26

sound complete

Question 27

sound

Answer 27

generates only functional dependencies

Question 28

Complete

Answer 28

generate all functional dependencies

Question 29

Union Rule

Answer 29

a gives b and a gives y then a gives BY

Question 30

decomposition rule

Answer 30

if a gives by then a gives b and a gives y

Question 31

pseudotransitivity rule

Answer 31

if a gives b and yb gives c then ay gives c

Question 32

Computing Closure of F

Answer 32

* apply reflexivity and augmentation rule * add dependencies in f1 and f2 only if they can combine through transivity

Question 33

test a is a superkey

Answer 33

compute closure of a and see if it contains all the attributes

Question 34

Canonical Cover

Answer 34

testing simplified set of functional dependencies that has the same closure as the set

Question 35

extraneous attributes

Answer 35

attribute can be removed without changing the closure of F

Question 36

Extraneous from the left makes it stronger but...

Answer 36

A gives C implies AB gives C but it cant be the otherway around

Question 37

Extraneous from the right makes it weaker

Answer 37

AB gives CD cannot imply that AB only gives D

Question 38

restriction

Answer 38

all functional dependencies include only attributes of R

Question 39

Dependency Preservation Time

Answer 39

it takes polynomial time instead of exponential

Question 40

Simplified Test

Answer 40

only check dependencies in set F for a violation

Question 41

However simplified test using only F is incorrect when...

Answer 41

testing a relation in decomposition of R

Question 42

for every attribute of A subgroup of R, check that closure of a includes...

Answer 42

no attributes of R-a or includes all attributes of R

Question 43

nothing