Database Development

Process

Dr. Shams Qazi

shams.qazi@seecs.edu.pk
Database Systems – CS220
ANSI/SPARC Architecture

 One of the major advantages of database systems is data

abstraction. Data abstraction is achieved by defining
different views of the data. Each view isolates higher-level
views from data representation details.
 The ANSI/SPARC architecture defined in 1975 consists of
three views:
 Internal View - The physical representation of the database on the
computer. How the data is stored.
 Conceptual View - The logical structure of the database that describes
what data is stored and its relationships.
 External View - The user's view of the database that provides the part
of the database relevant to the user.
ANSI/SPARC Architecture

Users

External External External External

view view view
Schema

3-schema
Conceptual
architecture
Conceptual
view
Schema

Internal
Internal view
Schema DBMS

DB
Benefits of 3-Schema Architecture

External Level: Each user can access the data, but have their own view of the
data independent of other users.
 Logical data independence - conceptual schema changes do not affect
external views.
Conceptual Level: Single shared data representation for all applications and
users which is independent of physical data storage.
 Users do not have to understand physical data representation details.
 The DBA can change the storage structures without affecting users or
applications. Physical data independence - conceptual schema not affected
by physical changes such as adding indexes or distributing data.
Internal (Physical) Level: Provides standard facilities for interacting with
operating system for space allocation and file manipulation.
MS Access and the 3-Schema Architecture

External Level: Microsoft Access does not call them views, but you can store
queries and use the results in other queries (like a view).
 External schema is the query (view) name and the attribute metadata.
Conceptual Level: All tables and field definitions are in the schema (accessible
from the Tables tab).
 Note that conceptual schema is not the data but the metadata.
Physical Level:
 Access represents all data in a single file whose layout it controls.
 The system processes this raw data file by knowing locations and offsets
of relations and fields.
Question?

Question: What are the three levels of views in the ANSI/SPARC

architecture starting with the view closest to the user?
A. Physical, Conceptual, External
B. External, Physical, Conceptual
C. Physical, External, Conceptual
D. External, Conceptual, Internal
E. User, Logical, System
Question?

Question: Assume you have a Java program accessing data stored in

a file. Select one true statement.
A. The file organization is changed. The physical view is where this
change is made.
B. A field is added to the database. The conceptual view is
changed.
C. A user account has restricted access to the file. The external view
must be changed.
D. More than one of the above
Relational Model
Relational Model - History

The relational model was proposed by Edgar Frank Codd in 1970.

One of the first relational database systems, System R, developed at IBM led
to several important breakthroughs:
 the first version of SQL
 various commercial products such as Oracle and DB2
 extensive research on concurrency control, transaction management,
and query processing and optimization

Commercial implementations (RDBMSs) appeared in the late 1970s and early

1980s. Currently, the relational model is the foundation of the majority of
commercial database systems.
Relational Model - Definitions

 A relation is a table with columns and rows.

 An attribute is a named column of a relation.

 A tuple is a row of a relation.

 A domain is a set of allowable values for one or more attributes.

 The degree of a relation is the number of attributes it contains.

 The cardinality of a relation is the number of tuples it contains.

 A relational database is a collection of normalized relations with distinct

relation names.
 The intension of a relation is the structure of the relation including its domains.

 The extension of a relation is the set of tuples currently in the relation.

Traditional File Processing Systems

relation
attributes

tuples

Degree = 7 Domain of Unit

Degree = 7
Cardinality = 77 Price is currency.
Cardinality = 77
Question?

 Question: Given the three definitions, select the ordering that contains their
related definitions.
1. relation
2. tuple
3. attribute

A. column, row, table

B. row, column, table
C. table, row, column
D. table, column, row
Question?

Question: A database table has 10 rows and 5 columns. Select one true statement.
A. The table's degree is 50.
B. The table's cardinality is 5.
C. The table's degree is 10.
D. The table's cardinality is 10.
Question?

 What is the name of the relation?

 What is the cardinality of the relation?
 What is the degree of the relation?
 What is the domain of order date? What is the domain of order id?
 What is larger the size of the intension or extension?
Relational Model Formal Definition

The relational model may be visualized as tables and fields, but it is formally
defined in terms of sets and set operations.

A relation schema R with attributes A =<A1, A2, …, An> is denoted R (A1, A2, …,
An) where each Ai is an attribute name that ranges over a domain Di denoted
dom(Ai).
Example: Product (id, name, supplierId, categoryId, price)
 R = Product (relation name)
 Set A = {id, name, supplierId, categoryId, price}
 dom(price) is set of all possible positive currency values
 dom(name) is set of all possible strings that represent people's names
Relation Schemas and Instances

A relation schema is a definition of a single relation.

 The relation schema is the intension of the relation.

A relational database schema is a set of relation schemas (modeling a particular

domain).

A relation instance denoted r(R) over a relation schema R(A1, A2, …, An) is a set of n-
tuples <d1, d2, ..., dn> where each di is an element of dom(Ai) or is null.
 The relation instance is the extension of the relation.
 A value of null represents a missing or unknown value.
Cartesian Product

The Cartesian product written as D1  D2 is a set operation that takes two sets D1 and D2 and
returns the set of all ordered pairs such that the first element is a member of D1 and the
second element is a member of D2. Example:
 D1 = {1,2,3}

 D2 = {A,B}
 D1 x D2 = {(1,A), (2,A), (3,A), (1,B), (2,B), (3,B)}
Practice Questions:
1) Compute D2 x D1.
2) Compute D2 x D2.
3) If |D| denotes the number of elements in set D, how many elements are there in D1 x D2
4) What is the cardinality of D1 x D2 x D1 x D1?
Relation Instance

A relation instance r(R) can also be defined as a subset of the Cartesian product of
the domains of all attributes in the relation schema. That is,
r(R) ⊆ dom(A1) x dom(A2) x … x dom(An)
Example:
 R = Person(id, firstName, lastName)
 dom(id) = {1,2}, dom(firstName) = {Joe, Steve}, dom(lastName) = {Jones, Perry}
 dom(id) x dom(firstName) x dom(lastName) =
 { (1,Joe,Jones), (1,Joe,Perry), (1,Steve,Jones), (1,Steve,Perry), (2,Joe,Jones),
(2,Joe,Perry), (2,Steve,Jones), (2,Steve,Perry)}
 Assume our DB stores people Joe Jones and Steve Perry, then
r(R) = { (1,Joe, Jones), (2,Steve,Perry)}.
Properties of Relation

1. Each relation name is unique. (No two relations have the same name.)
2. Each cell of the relation (value of a domain) contains exactly one atomic (single)
value.
3. Each attribute of a relation has a distinct name.
4. The values of an attribute are all from the same domain.
5. Each tuple is distinct. There are no duplicate tuples. (This is because relations are
sets. In SQL, relations are bags.)
6. The order of attributes is not really important.
 Note that this is different that a mathematical relation and our definitions which specify an
ordered tuple. The reason is that the attribute names represent the domain and can be
reordered.
7. The order of tuples has no significance.
 Relational Keys

o Keys are used to uniquely identify a tuple in a relation.

 Note that keys apply to the relational schema not to the relational instance. That is, looking at the current data
cannot tell you for sure if the set of attributes is a key.

o A superkey is a set of attributes that uniquely identifies a tuple in a relation.

o A key is a minimal set of attributes that uniquely identifies a tuple in a relation.

o A candidate key is one of the possible keys of a relation.
o A primary key is the candidate key designated as the distinguishing key of a relation.

o A foreign key is a set of attributes in one relation referring to the primary key of another
relation.

 Foreign keys allow referential integrity to be enforced.

Keys and Superkey Question

True or false: It is possible to have more than one key for a table and
the keys may have different numbers of attributes.

A. true

B. false
Keys and Superkey Question

True or false: It is possible to always determine if a field is a key by

looking at the data in the table.

A. true

B. false
Example Relations

Employee-Project Database:
 Employees have a unique number, name, title, and salary.
 Projects have a unique number, name, and budget.
 An employee may work on multiple projects and a project may have multiple
employees. An employee on a project has a particular responsibility and duration
on the project.
Relations:
 Emp (eno, ename, title, salary)

 Proj (pno, pname, budget)

 WorksOn (eno, pno, resp, dur)
Underlined attributes denote keys.
Example Relation Instances
Emp Relation WorksOn Relation
eno ename title salary eno pno resp dur
E1 J. Doe EE 30000 E1 P1 Manager 12
E2 M. Smith SA 50000 E2 P1 Analyst 24
E3 A. Lee ME 40000 E2 P2 Analyst 6
E4 J. Miller PR 20000 E3 P3 Consultant 10
E5 B. Casey SA 50000 E3 P4 Engineer 48
E6 L. Chu EE 30000 E4 P2 Programmer 18
E7 R. Davis ME 40000 E5 P2 Manager 24
E8 J. Jones SA 50000 E6 P4 Manager 48
E7 P3 Engineer 36
Proj Relation E7 P5 Engineer 23
pno pname budget E8 P3 Manager 40
P1 Instruments 150000
P2 DB Develop 135000 Questions:
P3 CAD/CAM 250000 1. Is ename a key for emp?
P4 Maintenance 310000 2. Is eno a key for WorksOn?
P5 CAD/CAM 500000 3. List all the superkeys for WorksOn.
Practice Questions

Consider a relation storing driver information including:

 SSN, name, driver's license number and state (unique together)

SSN Name LicNum LicState Questions:

123-45-6789 S. Smith 123-456 IA 1) List the candidate keys for the
relation.
111-11-1111 A. Lee 123-456 NY 2) Pick a primary key for the relation.
222-22-2222 J. Miller 555-111 MT 3) Is name a candidate key for
333-33-3333 B. Casey 678-123 OH Person?
4) List all the superkeys for Person
444-44-4444 A. Adler 456-345 IA
Assumptions:
1) A person has only one driver’s license.
2) A driver’s license uniquely
identifies a person.
Relational Integrity

Integrity rules are used to insure the data is accurate.

Constraints are rules or restrictions that apply to the database and limit the data values it may
store.
Types of constraints:

Domain constraint - Every value for an attribute must be an element of the attribute's
domain or be null.
 Null represents a value that is currently unknown or not applicable.
 null is not the same as zero or an empty string.
Entity integrity constraint - In a base relation, no attribute of a primary key can be null.
Referential integrity constraint - If a foreign key exists in a relation, then the foreign key value
must match a primary key value of a tuple in the referenced relation or be null.
Foreign Key Example
Emp Relation WorksOn Relation
eno ename title salary eno pno resp dur
E1 J. Doe EE 30000 WorksOn.eno is E1 P1 Manager 12
E2 M. Smith SA 50000 FK to Emp.eno E2 P1 Analyst 24
E3 A. Lee ME 40000 E2 P2 Analyst 6
E4 J. Miller PR 20000 E3 P3 Consultant 10
E5 B. Casey SA 50000 E3 P4 Engineer 48
WorksOn.pno is E4 P2 Programmer 18
E6 L. Chu EE 30000
FK to Proj.pno
E7 R. Davis ME 40000 E5 P2 Manager 24
E8 J. Jones SA 50000 E6 P4 Manager 48
E7 P3 Engineer 36
Proj Relation E7 P5 Engineer 23
pno pname budget E8 P3 Manager 40
P1 Instruments 150000
P2 DB Develop 135000
P3 CAD/CAM 250000
P4 Maintenance 310000
P5 CAD/CAM 500000
Integrity Constraint Question

Question: What constraint says that a primary key field cannot be null?

A. domain constraint

B. referential integrity constraint

C. entity integrity constraint

Entity Integrity Constraint Question

Question: A primary key has three fields. Only one field is null. Is the
entity integrity constraint violated?

A. Yes

B. No
Referential Integrity Constraint Question

Question: A foreign key has a null value in the table that contains the
foreign key fields. Is the referential integrity constraint violated?

A. Yes

B. No
Integrity Question
Emp Relation WorksOn Relation
eno ename title salary eno pno resp dur
E1 J. Doe EE AS E1 P0 null 12
E2 null SA 50000 E2 P1 Analyst null
E3 A. Lee 12 40000 null P2 Analyst 6
E4 J. Miller PR 20000 E3 P3 Consultant 10
E5 B. Casey SA 50000 E9 P4 Engineer 48
null L. Chu EE 30000 E4 P2 Programmer 18
E7 R. Davis ME null E5 null Manager 24
E8 J. Jones SA 50000 E6 P4 Manager 48
E7 P6 Engineer 36
Proj Relation E7 P4 Engineer 23
pno pname budget null null Manager 40
P1 Instruments 150000
P2 DB Develop 135000
Question:
P3 CAD/CAM 250000
How many violations of integrity constraints?
P4 Maintenance 310000
P5 null null
General Constraints

There are more general constraints that some DBMSs can enforce. These
constraints are often called enterprise constraints or semantic integrity constraints.

Examples:
 An employee cannot work on more than 2 projects.
 An employee cannot make more money than their manager.
 An employee must be assigned to at least one project.

Ensuring the database follows these constraints is usually achieved using triggers.
Relational Algebra

A query language is used to update and retrieve data that is stored in a data model.

Relational algebra is a set of relational operations for retrieving data.

 Just like algebra with numbers, relational algebra consists of operands (which are
relations) and a set of operators.

Every relational operator takes as input one or more relations and produces a relation as
output.

 Closure property - input is relations, output is relations

 Unary operations - operate on one relation

 Binary operations - have two relations as input

A sequence of relational algebra operators is called a relational algebra expression.

Relational Algebra Operators

Relational Operators:
 Selection σ
 Projection Π
 Cartesian product ×
 Join
 Union ∪
 Difference -
 Intersection ∩

Note that relational algebra is the foundation of ALL relational database

systems. SQL gets translated into relational algebra.
Selection Operator

The selection operation is a unary operation that takes in a relation as input and
returns a new relation as output that contains a subset of the tuples of the input
relation.
 That is, the output relation has the same number of columns as the input relation,
but may have less rows.

To determine which tuples are in the output, the selection operation has a
specified condition, called a predicate, that tuples must satisfy to be in the
output.
 The predicate is similar to a condition in an if statement.
Selection Operation – Formal Definition

The selection operation on relation R with predicate F is denoted by σF(R).

σF(R)={t | t∈R and F(t) is true}
where
 R is a relation, t is a tuple variable
 F is a formula (predicate) consisting of
 operands that are constants or attributes
 comparison operators: <, >, =, ≠, ≤, ≥
 logical operators: AND, OR, NOT
Selection Example

Emp Relation σtitle = 'EE' (Emp)

eno ename title salary
E1 J. Doe EE 30000 eno ename title salary
E2 M. Smith SA 50000 E1 J. Doe EE 30000
E3 A. Lee ME 40000 E6 L. Chu EE 30000
E4 J. Miller PR 20000
E5 B. Casey SA 50000 σsalary > 35000 OR title = 'PR' (Emp)
E6 L. Chu EE 30000
eno ename title salary
E7 R. Davis ME 40000
E8 J. Jones SA 50000 E2 M. Smith SA 50000
E3 A. Lee ME 40000
E4 J. Miller PR 20000
E5 B. Casey SA 50000
E7 R. Davis ME 40000
E8 J. Jones SA 50000
Selection Question?

Question: Given this table and the query:

σsalary > 50000 or title='PR'(Emp)

How many rows are returned? Emp Relation

eno ename title salary
E1 J. Doe EE 30000
A. 0 E2 M. Smith SA 50000
E3 A. Lee ME 40000
B. 1 E4 J. Miller PR 20000
C. 2 E5 B. Casey SA 50000
E6 L. Chu EE 30000
D. 3 E7 R. Davis ME 40000
E8 J. Jones SA 50000
Selection Question (2)

Question: Given this table and the query:

σsalary > 50000 or title='PR'(Emp)

How many columns are returned? Emp Relation

eno ename title salary
E1 J. Doe EE 30000
A. 0 E2 M. Smith SA 50000
E3 A. Lee ME 40000
B. 2 E4 J. Miller PR 20000
C. 3 E5 B. Casey SA 50000
E6 L. Chu EE 30000
D. 4 E7 R. Davis ME 40000
E8 J. Jones SA 50000
Selection Questions

WorksOn Relation
eno pno resp dur Write the relational algebra expression that:
E1 P1 Manager 12 1. Returns all rows with an employee working
E2 P1 Analyst 24 on project P2.
E2 P2 Analyst 6 2. Returns all rows with an employee who is
E3 P3 Consultant 10 working as a manager on a project.
E3 P4 Engineer 48 3. Returns all rows with an employee working as
E4 P2 Programmer 18 a manager for more than 40 months.
E5 P2 Manager 24
E6 P4 Manager 48 Show the resulting relation for each case.
E7 P3 Engineer 36
E7 P5 Engineer 23
E8 P3 Manager 40
Projection Operator

The projection operation is a unary operation that takes in a relation as input and
returns a new relation as output that contains a subset of the attributes of the
input relation and all non-duplicate tuples.
 The output relation has the same number of tuples as the input relation unless
removing the attributes caused duplicates to be present.
 Question: When are we guaranteed to never have duplicates when performing
a projection operation?

Besides the relation, the projection operation takes as input the names of the
attributes that are to be in the output relation.
Projection Operation Formal Definition

The projection operation on relation R with output attributes

A1,…,A m is denoted by ΠA ,…,A (R).
1 m

ΠA1,…,Am(R)={t[A1,…, A m] | t∈R}

where
R is a relation, t is a tuple variable
 {A1,…,Am} is a subset of the attributes of R over which the projection will
be performed.
 Order of A1,…, Am is significant in the result.
 Cardinality of Π A ,…,A (R) is not necessarily the same as R because
1 m
of duplicate removal.
Projection Example

Emp Relation Π eno,ename (Emp) Π title (Emp)

eno ename title salary eno ename title
E1 J. Doe EE 30000 E1 J. Doe EE
E2 M. Smith SA 50000 E2 M. Smith SA
E3 A. Lee ME 40000 E3 A. Lee ME
E4 J. Miller PR 20000 E4 J. Miller PR
E5 B. Casey SA 50000 E5 B. Casey
E6 L. Chu EE 30000 E6 L. Chu
E7 R. Davis ME 40000 E7 R. Davis
E8 J. Jones SA 50000 E8 J. Jones
Projection Question

Question: Given this table and the query:

Π title (Emp)

How many rows are returned? Emp Relation

eno ename title salary
E1 J. Doe EE 30000
A. 0 E2 M. Smith SA 50000
E3 A. Lee ME 40000
B. 2 E4 J. Miller PR 20000
C. 4 E5 B. Casey SA 50000
E6 L. Chu EE 30000
D. 8 E7 R. Davis ME 40000
E8 J. Jones SA 50000
 Projection Questions

WorksOn Relation
eno pno resp dur
E1 P1 Manager 12 Write the relational algebra expression that:
E2 P1 Analyst 24
1. Returns only attributes resp and dur.
E2 P2 Analyst 6
2. Returns only eno.
E3 P3 Consultant 10
3. Returns only pno.
E3 P4 Engineer 48
E4 P2 Programmer 18
E5 P2 Manager 24 Show the resulting relation for each case.
E6 P4 Manager 48
E7 P3 Engineer 36
E7 P5 Engineer 23
E8 P3 Manager 40

Page 40