Schema Refinement

The Schema Refinement refers to refine the schema by using some technique. The best technique of schema
refinement is decomposition. Normalisation or Schema Refinement is a technique of organizing the data in
the database. It is a systematic approach of decomposing tables to eliminate data redundancy and undesirable
characteristics like Insertion, Update and Deletion Anomalies.
Redundancy refers to repetition of same data or duplicate copies of same data stored in different locations.
Anomalies refers to the problems occurred after poorly planned and normalised databases where all the data
is stored in one table which is sometimes called a flat file database.
Anomalies or problems facing without normalization (problems due to redundancy)
Anomalies refers to the problems occurred after poorly planned and un-normalised databases where all the
data is stored in one table which is sometimes called a flat file database. Let us consider such type of schema

Here all the data is stored in a single table which causes redundancy of data or say anomalies as SID and
Sname are repeated once for same CID. Let us discuss anomalies one by one.
Due to redundancy of data we may get the following problems, those are
Insertion anomalies: It may not be possible to store some information unless some other information is
stored as well.
Redundant storage: some information is stored repeatedly.
Update anomalies: If one copy of redundant data is updated, then inconsistency is created unless all
redundant copies of data are updated.
Deletion anomalies: It may not be possible to delete some information without losing some other information
as well.
Problem in updation / updation anomaly – If there is updation in the fee from 5000 to 7000, then we have to
update FEE column in all the rows, else data will become inconsistent.
Insertion Anomaly and Deletion Anomaly: These anomalies exist only due to redundancy, otherwise they
do not exist.
Insertion Anomalies: New course is introduced C4, But no student is there who is having C4 subject.

Because of insertion of some data, It is forced to insert some other dummy data.
Deletion Anomaly
Deletion of S3 student cause the deletion of course.
Because of deletion of some data forced to delete some other useful data.
Solutions to Anomalies: Decomposition of Tables – Schema Refinement

There are some Anomalies in this again

What is the Solution?
Solution: decomposing into relations as shown below

To Avoid Redundancy and problems due to redundancy, we use refinement technique called decomposition.
Decomposition: Process of decomposing a larger relation into smaller relations.
Each of smaller relations contain subset of attributes of original relation.
Functional dependencies:
Functional dependency is a relationship that exist when one attribute uniquely determines another attribute.
Functional dependency is a form of integrity constraint that can identify schema with redundant storage
problems and to suggest refinement.
A functional dependency A→B in a relation holds true if two tuples having the same value of attribute A
also have the same value of attribute B
IF t1.X=t2.X then t1.Y=t2.Y where t1, t2 are tuples and X, Y are attributes.
Reasoning about functional dependencies
Armstrong Axioms
Armstrong axioms defines the set of rules for reasoning about functional dependencies and also to infer all
the functional dependencies on a relational database.
Various axioms rules or inference rules
Primary axioms:

Secondary or derived axioms

Attribute closure: Attribute closure of an attribute set can be defined as set of attributes which can be
functionally determined from it.
NOTE: To find attribute closure of an attribute set
 Add elements of attribute set to the result set.
 Recursively add elements to the result set which can be functionally determined from the elements of
result set.

Types of functional dependencies

 Trivial functional dependency:-If X→Y is a functional dependency where Y subset X, these type of
FD’s called as trivial functional dependency.
 Non-trivial functional dependency:-If X→Y and Y is not subset of X then it is called non-trivial
functional dependency.
 Completely non-trivial functional dependency:-If X→Y and X∩Y=Ф(null) then it is called
completely non-trivial functional dependency.
Prime and non-prime attributes
Attributes which are parts of any candidate key of relation are called as prime attribute, others are non-prime
attributes.
Candidate Key is minimal set of attributes of a relation which can be used to identify a tuple uniquely.
Consider student table: student (sno, sname,sphone,age)
We can take sno as candidate key. We can have more than 1 candidate key in a table. Types of candidate
keys:
Simple (having only one attribute)
Composite (having multiple attributes as candidate key) Super Key
Super Key is set of attributes of a relation which can be used to identify a tuple uniquely.
 Adding zero or more attributes to candidate key generates super key.
 A candidate key is a super key but vice versa is not true.
Consider student table: student (sno, sname,sphone,age)
We can take sno, (sno, sname) as super key
Finding candidate keys problems:
Normalization
Normalization is a process of designing a consistent database with minimum redundancy which support data
integrity by grating or decomposing given relation into smaller relations preserving constraints on the
relation. Normalisation removes data redundancy and it will helps in designing a good data base which
involves a set of normal forms as follows -
1) First normal form (1NF)
2) Second normal form (2NF)
3) Third normal form (3NF)
4) Boyce coded normal form (BCNF)
5) Forth normal form (4NF)
6) Fifth normal form (5NF)
7) Sixth normal form (6NF)
8) Domain key normal form.
1) First normal form: A relation is said to be in first normal form if it contains all atomic values or single
values.
Example:
Domain Courses
Programming C , java
Web designing HTML , PHP
The above table consist of multiple values in single columns which can be reduced into atomic values by
using first normal form as follows-
Domain Courses
Programming C
Programming Java
Web designing HTML
Web designing PHP
2) Second normal form: A relation is said to be in second normal form if it is in first normal form without
any partial dependencies.
In second normal form non-prime attributes should not depend on proper subset of key attributes. Example:
Student id Student name Project Id Project name

Here (student id, project id) are key attributes and (student name, project name) are non-prime attributes. It
is decomposed as-
Student id Student name Project id
 Project id Project name

3) Third normal form: A relation is said to be in third normal form , if it is already in second normal form
and no transitive dependencies exists.
Transitive dependency – If A->B and B->C are two FDs then A->C is called transitive dependency.
A relation is in 3NF if at least one of the following condition holds in every non-trivial function dependency
X –> Y
 X is a super key.
 Y is a prime attribute (each element of Y is part of some candidate key).

Student id Student name City country ZIP

It is decomposed as
Student id Student name ZIP

ZIP city country

4) Boyce normal form: It is an extension of third normal form where in a functional dependency X→A, X
must be a super key.
A relation is in BCNF if in every non-trivial functional dependency X –> Y, X is a super key.
5) Fourth normal form: A relation is said to be in fourth normal form if it is in third normal form and no
multi value dependencies should exist between attributes.
Note: In some cases multi value dependencies may exist not more than one time in a given relation.
6) Fifth normal form: fifth normal form is related to join dependencies.
A relation R is said to be in fifth normal form if for every join dependency JD join {R1 , R2 ,…….RN} that
holds over relation R one of the following statements must be true
 Ri =R for some i
 the join dependency is implied by the set of those functional dependency over relation R in which the
left side is key attribute for R.
NOTE: if the relation schema is a third normal form and each of its keys consist of single attribute, we can
say that it can also be in fifth normal form.
A join dependency JD join {R1, R2, ……RN} is said to hold for a relation R if R1,R2…..RN this
decomposition is a loss less join decomposition of R.
When a relation is in forth normal form and decompose further to eliminate redundancy and anomalies due
to insert or update or delete operation, there should not be any loss of data or should not create a new record
when the decompose tables are rejoin.
7) Domain key normal form: A domain key normal form keeps a constraint that every constraint on the
relation is a logical sequence of definition of keys and domains.
8) Sixth normal form: A relation is said to be in sixth normal form such that the relation R should not contain
any non-trivial join dependencies.
Also sixth normal form considers temporal dimensions(time) to the relational model.
Key Points related to normal forms
1. BCNF is free from redundancy.
2. If a relation is in BCNF, then 3NF is also also satisfied.
3. If all attributes of relation are prime attribute, then the relation is always in 3NF.
4. A relation in a Relational Database is always and at least in 1NF form.
5. Every Binary Relation ( a Relation with only 2 attributes ) is always in BCNF.
6. If a Relation has only singleton candidate keys( i.e. every candidate key consists of only 1 attribute),
then the Relation is always in 2NF( because no Partial functional dependency possible).
7. Sometimes going for BCNF form may not preserve functional dependency. In that case go for BCNF
only if the lost FD(s) is not required, else normalize till 3NF only.
8. There are many more Normal forms that exist after BCNF, like 4NF and more. But in real world
database systems it’s generally not required to go beyond BCNF.
Problems on normal forms:
Problem 1:
Find the highest normal form in R (A, B, C, D, E) under following functional dependencies. ABC --> D
CD --> AE
Solution:
Important Points for solving above type of question.
 It is always a good idea to start checking from BCNF, then 3 NF and so on.
  If any functional dependency satisfied a normal form then there is no need to check for lower normal
form. For example, ABC –> D is in BCNF (Note that ABC is a super key), so no need to check this
dependency for lower normal forms.
Candidate keys in given relation are {ABC, BCD}
BCNF: ABC -> D is in BCNF. Let us check CD -> AE, CD is not a super key so this dependency is not in
BCNF. So, R is not in BCNF.
3NF: ABC->D we don’t need to check for this dependency as it already satisfied BCNF. Let us consider CD-
>AE. Since E is not a prime attribute, so relation is not in 3NF.
2NF: In 2NF, we need to check for partial dependency. CD which is a proper subset of a candidate key and
it determine E, which is non-prime attribute. So, given relation is also not in 2 NF. So, the highest normal
form is 1 NF.

Problem 2:
Find the highest normal form of a relation R (A, B, C, D, E) with
FD set as
{BC->D,
AC->BE,
B->E}
Step 1: As we can see, (AC) += {A, C, B, E, D} but none of its subset can determine all attribute of relation,
so AC will be candidate key. A or C can’t be derived from any other attribute of the relation, so there will be
only 1 candidate key {AC}.
Step 2: Prime attribute are those attribute which are part of candidate key {A,C} in this example and others
will be non-prime {B,D,E} in this example.
Step 3: The relation R is in 1st normal form as a relational DBMS does not allow multi-valued or composite
attribute.
The relation is in 2nd normal form because BC->D is in 2nd normal form (BC is not proper subset of
candidate key AC) and AC->BE is in 2nd normal form (AC is candidate key) and B->E is in 2nd normal
form (B is not a proper subset of candidate key AC).
The relation is not in 3rd normal form because in BC->D (neither BC is a super key nor D is a prime attribute)
and in B->E (neither B is a super key nor E is a prime attribute) but to satisfy 3rd normal for, either LHS of
an FD should be super key or RHS should be prime attribute.
So the highest normal form of relation will be 2nd Normal form.
Decomposition: It is the process of splitting original table into smaller relations such that attribute sets of
two relations will be the subset of attribute set of original table.
Rules of decomposition:
If ‘R’ is a relation splitted into ‘R1’ and ‘R2’ relations, the decomposition done should satisfy following-
1) Union of two smaller subsets of attributes gives all attributes of ‘R’.
R1 (attributes) U R2 (attributes) =R (attributes)
2) Both relations interaction should not give null value.
 R1 (attributes) ∩ R2 (attributes)! =null
3) Both relations interaction should give key attribute.
R1 (attribute) ∩ R2 (attribute) =R (key attribute)
Properties of decomposition:
Lossless decomposition: while joining two smaller tables no data should be lost and should satisfy all the
rules of decomposition. No additional data should be generated on natural join of decomposed tables.
Example 2 for lossless decomposition:

Lossy join decomposition: if information is lost after joining and if do not satisfy any one of the above rules
of decomposition.
Example 1:
Example 2:

In above examples, on joining decomposed tables, extra tuples are generated. so it is lossy join
decomposition.
Dependency preservation: functional dependencies should be satisfied even after splitting relations and
they should be satisfied by any of splitted tables.
Dependency Preservation
A Decomposition D = {R1, R2, R3….Rn} of R is dependency preserving wrt a set F of Functional
dependency if
(F1 F2 … Fm)+ = F+.
Consider a relation R
R ---> F{...with some functional dependency(FD)....}

R is decomposed or divided into R1 with FD { f1 } and R2 with { f2 }, then there can be three cases:
f1 U f2 = F -----> Decomposition is dependency preserving. f1 U f2 is a subset of F -----> Not Dependency
preserving. f1 U f2 is a super set of F -----> This case is not possible.
Example for dependency preservation:

Lack of redundancy: It is also known as repetition of information. The proper decomposition should not
suffer from any data redundancy.