Data modeling using the entity-relationship model
Outline: Data modeling using ER-model
(Chapter 3 - 3rd, 4th, 5th ed.)
• Data modeling
• Entity-relationship model
- Entity types
- strong entities
- weak entities
- Relationships among entities
- Attributes - attribute classification
- Constraints
- cardinality constraints
- participation constraints
• ER-to-Relation-mapping
Sept. 2012
Yangjun Chen
ACS-3902
1
Data modeling using the entity-relationship model
Entity-relationship model (ER model)
P. 41-2: “This model and its variations are frequently used
for the conceptual design of database applications, and many
database design tools employ its concepts.”
ER model:
• is used to create a conceptual data model that reflects all the
user data requirements.
•It includes detailed descriptions of
•entity types,
•relationships, and
•constraints
•no implementation details so can be used for communication
with non-technical users
Sept. 2012
Yangjun Chen
ACS-3902
2
Data modeling using the entity-relationship model
Example
The company database keeps track of a company’s
employees, departments, and projects:
Requirements:
concerning the department:
1.
2.
3.
4.
5.
company is organized into departments
a department has a unique name, a unique number, and a specific
employee is its’ manager
we track the start date for the manager function
a department may be in several locations
a department controls a number of projects
concerning the project:
6.
a project has a unique name, a unique number, and is in a single
location
Sept. 2012
Yangjun Chen
ACS-3902
3
Data modeling using the entity-relationship model
example continued
concerning the employee:
7.
each employee has a name, social insurance number,
address, salary, sex, and birth date
8. an employee is assigned to one department but may work on several
projects which are not necessarily controlled by the same department
9. we track the number of hours per week that an employee works on
each project
10. we keep track of the direct supervisor of each employee
11. we track the dependents of each employee (for insurance purposes)
concerning the dependent:
12. we record each dependent’s first name, sex, birth date,
and relationship to the employee
Sept. 2012
Yangjun Chen
ACS-3902
4
Data modeling using the entity-relationship model
The entities:
employee
dependent
department
project
Sept. 2012
Yangjun Chen
ACS-3902
5
Data modeling using the entity-relationship model
The entities:
fname
minit
lname
name
sex
ssn
address
dependent
salary
name
employee
sex
birthdate
relationship
bdate
name number
location
department
Sept. 2012
Yangjun Chen
project
name number
ACS-3902
location
6
Data modeling using the entity-relationship model
With relationships:
1
N
department
works for
1
1
employee
supervisee
manages
1
supervision
controls
N
supervisor
N
1
N
M
works on
1
project
dependents of
N
partial constraint
total constraint
dependent
Sept. 2012
Yangjun Chen
ACS-3902
7
Data modeling using the entity-relationship model
name number
With attributes, etc:
fname
name
ssn
bdate
degree
1
lname
minit
sex
1
startdate
employee
supervisor
department
works for
N
salary
address
1
location
1
controls
manages
1
N hours
N
supervisee
number of
employees
N
M
1
project
works on
supervision
dependents of
name number
location
N
Figure 3.2
dependent
name
Sept. 2012
sex
birthdate relationship
Yangjun Chen
ACS-3902
P.46
8
Data modeling using the entity-relationship model
Let’s consider a university environment:
At a high level, we need to:
•offer courses
•register students
•assign instructors to courses
•assign grades
•anything else?
Sept. 2012
Yangjun Chen
ACS-3902
9
Data modeling using the entity-relationship model
Concepts (you should understand)
Attribute classification
•null values
•simple attribute
•multivalued attribute
not applicable,
unknown, missing
•composite attribute
•complex (nested) attributes
•key attribute
•domain
Attribute storage
•stored & derived attribute
Sept. 2012
Yangjun Chen
ACS-3902
10
Data modeling using the entity-relationship model
Concepts (you should understand)
Relationships
•degree of a relationship
•recursive relationship
•role names
•constraints
cardinality: m-n, 1-n, 1-1
participation: partial, total (existence dependency)
Sept. 2012
Yangjun Chen
ACS-3902
11
Data modeling using the entity-relationship model
Entities
•entity type
•strong entity
- key attribute
- entity has a key attribute or a combination of attributes
which can be used as a key.
•weak entity
No key attributes. Entities belonging to a weak entity type are
identified by being related to specific entities from another entity type
in combination with some of their attribute values.
- identifying owner
- identifying relationship
- partial key
Sept. 2012
Yangjun Chen
ACS-3902
12
Data modeling using the entity-relationship model
Instructors: let’s assume this classification includes
instructors, professors, part-time people (at least for
now). These people have SINs, employee numbers,
names, addresses, offices, phones, ...
employee no
SIN
name
address
instructor
office
degree
Sept. 2012
phone
Yangjun Chen
ACS-3902
13
Data modeling using the entity-relationship model
Is there a key attribute? What are the domains? Can any attribute be
null? Is any attribute composite, derived, complex, multivalued?
employee no
SIN
name
address
instructor
office
degree
phone
Is this a weak entity or a strong entity?
Should department be an attribute?
Sept. 2012
Yangjun Chen
ACS-3902
14
Data modeling using the entity-relationship model
Departments: obviously instructors are employed by the
University and associated with a department
A department has a name, number, office, chair, ...
name
Dnumber
office
department
phone
chair
Sept. 2012
Yangjun Chen
ACS-3902
15
Data modeling using the entity-relationship model
Is there a key attribute? What are the domains? Can any attribute be
null? Is any attribute composite, derived, complex, multivalued?
name
Dnumber
office
department
phone
chair
Should chair be an attribute, or is there a relationship
between two entity types?
Sept. 2012
Yangjun Chen
ACS-3902
16
Data modeling using the entity-relationship model
Employs relationship: If we assume the relationship
between department and instructor is 1:N then we only
associate each department with a single instructor, but we
associate any number of instructors with a single department
department
1
employs
N
instructor
1:N is the cardinality of the relationship
the relationship is of degree 2; it is a binary relationship
Both entities are considered strong entities
Sept. 2012
Yangjun Chen
ACS-3902
17
Data modeling using the entity-relationship model
Consider some instances
department
employs
d1
e1
e2
e3
e4
e5
e6
e7
d2
d3
Sept. 2012
instructor
Yangjun Chen
ACS-3902
18
Data modeling using the entity-relationship model
Chair relationship: A department has a chair who has
special responsibilities. One person (instructor) is designated
as such.
department
1
1
chair
instructor
1:1 is the cardinality of the relationship
the relationship is of degree 2; it is a binary relationship.
Sept. 2012
Yangjun Chen
ACS-3902
19
Data modeling using the entity-relationship model
Continuing the discussion on data modeling
•Weak entity again
•Data analysis (at entity instance level)
•Mapping ER to relational database schema
•ER model comparison
Sept. 2012
Yangjun Chen
ACS-3902
20
Data modeling using the entity-relationship model
Weak entity types
a weak entity does not have a key of its own - may have a
partial key
the identifying relationship will have total participation for
the weak entity
e.g. consider courses and sections at UWinnipeg
Sept. 2012
Yangjun Chen
ACS-3902
21
Data modeling using the entity-relationship model
Consider courses and course sections
In the fall and winter we have:
91.1453/3-001 F Intro Computers staff MW 16:30-17:45 3C13 ..
91.1453/3-002 W Intro Computers staff MW 16:30-17:45 3C13 ..
91.1453/3-050 F Intro Computers staff T
18:00-21:00 3C13 ..
91.1453/3-051 W Intro Computers staff T
18:00-21:00 3C13 ..
Section numbers are 001, 002, 050, 051, …
Sections have a section number, a term, days and times, …
Sept. 2012
Yangjun Chen
ACS-3902
22
Data modeling using the entity-relationship model
Consider courses and course sections
name
term
course no
Section no
course
1
Offered-in
N
section
description
credit hours
meeting
Section is a weak entity - it has a discriminator (partial key), section number.
Section totally participates in the offered in relationship
PK (primary key) of Section is … (offered_in is an identifying relationship)
Is meeting multivalued?
Sept. 2012
Yangjun Chen
ACS-3902
23
Data modeling using the entity-relationship model
Data analysis:
instructor
m
teaches
n
course
n
m
textbook
uses
Note that teaches and uses are both binary relationships:
•we expect situations where a specific Instructor teaches a
specific Course, and where
•a specific Course uses a specific text
Sept. 2012
Yangjun Chen
ACS-3902
24
Data modeling using the entity-relationship model
Consider instances
instructor
Sept. 2012
course
Yangjun Chen
textbook
ACS-3902
25
Data modeling using the entity-relationship model
instructor
course
Z
Intro to X
jones
Intro to Y
smith
Advanced X
Courses offered are
Intro to X,
Intro to Y,
Z,
Advanced X,
Advanced Y
Advanced Y
Sept. 2012
Yangjun Chen
ACS-3902
26
Data modeling using the entity-relationship model
Consider instances of Instructor teaches Course
instructor
teaches
course
Z
Intro to X
jones
•Jones teaches Intro to X
Intro to Y
smith
Advanced X
Advanced Y
Sept. 2012
Yangjun Chen
ACS-3902
27
Data modeling using the entity-relationship model
instructor
teaches
course
Z
Intro to X
jones
•Jones teaches Intro to X
Intro to Y
smith
Advanced X
Advanced Y
There is a relationship between Instructor Jones and Course Intro to X
Sept. 2012
Yangjun Chen
ACS-3902
28
Data modeling using the entity-relationship model
instructor
teaches
course
Z
Intro to X
jones
•Jones teaches Intro to X
Intro to Y
smith
Advanced X
Advanced Y
This line connects the three dots
Sept. 2012
Yangjun Chen
ACS-3902
29
Data modeling using the entity-relationship model
instructor
teaches
course
Z
Intro to X
jones
Intro to Y
smith
•Jones teaches Intro to X
•Jones teaches Advanced X
Advanced X
Advanced Y
Sept. 2012
Yangjun Chen
ACS-3902
30
Data modeling using the entity-relationship model
instructor
teaches
course
Z
Intro to X
jones
Intro to Y
•Jones teaches Intro to X
•Jones teaches Advanced X
•Smith teaches Intro to Y
smith
Advanced X
Advanced Y
Sept. 2012
Yangjun Chen
ACS-3902
31
Data modeling using the entity-relationship model
instructor
teaches
course
Z
Intro to X
jones
Intro to Y
•Jones teaches Intro to X
•Jones teaches Advanced X
•Smith teaches Intro to Y
smith
Advanced X
•Smith teaches Advanced Y
Advanced Y
Sept. 2012
Yangjun Chen
ACS-3902
32
Data modeling using the entity-relationship model
instructor
teaches
course
Z
Intro to X
jones
Intro to Y
•Jones teaches Advanced X
•Smith teaches Intro to Y
smith
Advanced X
Advanced Y
Sept. 2012
•Jones teaches Intro to X
Yangjun Chen
ACS-3902
•Smith teaches Advanced Y
•Smith and Jones teach Z
together
There are two relationships:
one between Jones and Z; the
other between Smith and Z
33
Data modeling using the entity-relationship model
Now let us examine Course uses Textbook
instructor
teaches
course
uses
Z
Intro to X
jones
textbook
UML
distilled
Intro to Y
smith
Advanced X
Advanced Y
The mythical
man-month
Suppose we have two textbooks: The mythical
man-month, and UML distilled
Sept. 2012
Yangjun Chen
ACS-3902
34
Data modeling using the entity-relationship model
instructor
teaches
course
uses
Z
Intro to X
jones
textbook
UML
distilled
Intro to Y
smith
Advanced X
The mythical
man-month
Advanced Y
Intro to Y uses The mythical man-month
Sept. 2012
Yangjun Chen
ACS-3902
35
Data modeling using the entity-relationship model
instructor
teaches
course
uses
Z
Intro to X
jones
textbook
UML
distilled
Intro to Y
smith
Advanced X
The mythical
man-month
Advanced Y
Z uses both texts
Sept. 2012
Yangjun Chen
ACS-3902
36
Data modeling using the entity-relationship model
Let’s jump ahead to see what relations we eventually create
Section 9.1: ER-to-Relational mapping
1. Create a relation for each strong entity type
2. Create a relation for each weak entity type
•include primary key of owner (an FK - foreign key)
•owner’s PK + partial key becomes PK
3. For each binary 1:1 relationship choose an entity and
include the other’s PK in it as an FK. Include any
attributes of the relationship
4. For each binary 1:n relationship, choose the n-side entity
and include an FK w.r.t the other entity. Include any
attributes of the relationship
Sept. 2012
Yangjun Chen
ACS-3902
37
Data modeling using the entity-relationship model
5. For each binary M:N relationship, create a relation for the
relationship
•include PKs of both participating entities and any attributes of
the relationship
•PK is the concatenation of the participating entity PKs
6. For each multivalued attribute create a new relation
•include the PK attributes of the entity type
•PK is the PK of the entity type and the multivalued attribute
7. For each n-ary relationship, create a relation for the
relationship
•include PKs of all participating entities and any attributes of
the relationship
•PK may be the concatenation of the participating entity PKs
Sept. 2012
Yangjun Chen
ACS-3902
38
Data modeling using the entity-relationship model
1
N
department
works for
1
1
employee
supervisee
1
supervision
1
controls
N
supervisor
N
manages
N
M
works on
1
project
dependents of
N
dependent
Sept. 2012
Yangjun Chen
ACS-3902
39
Data modeling using the entity-relationship model
Consider the following requirements (very typical situation)
A company may be composed of Divisions, each of which may be
composed of Departments, each of which may be composed of
Groups.
We could model this as follows:
group
n
dep composed of
1
department
n
1
div composed of
division
n
c composed of
Sept. 2012
Yangjun Chen
ACS-3902
1
company
40
Data modeling using the entity-relationship model
It fits the requirements, but
•it is not flexible
•it does not accommodate any new groupings
•it does not allow the company to grow/change
•it is very rigid => it is a static model
group
n
dep composed of
1
department
n
1
div composed of
division
n
c composed of
Sept. 2012
Yangjun Chen
ACS-3902
1
company
41
Data modeling using the entity-relationship model
Consider the following model
•it is flexible
•it does accommodate any new groupings
•it does allow the company to grow/change
•it is not rigid => it is a dynamic model
•but not all constraints are implied in the model => we would
need supporting application code. What are these constraints?
composed of
m
n
higher_level
higher_level
Sept. 2012
Yangjun Chen
lower_level
ACS-3902
organization
42
Data modeling using the entity-relationship model
Consider the following model
•it is flexible, it does accommodate any new groupings, it does
allow the company to grow/change, it is a dynamic model again
•not all constraints are implied in the model => we would need
supporting application code. What are these constraints?
Constraint-des
n
composed of
org composed of
m
higher_
level
Sept. 2012
m
lower_level
organization
type
n lower_level
higher_level
organization
m
Yangjun Chen
example of
ACS-3902
n
43
Data modeling using the entity-relationship model
organization
org-name
Dell
sell-division
... ...
Composed-of
sub-org
org
Sell-division
Dell
Dell
Production-division
... ...
... ...
... ...
... ...
Example-of
org
org-type
company
Dell
sell-division division
... ...
organization-type
org-type
company
division
... ...
org-composed-of
sub-org-type
org-type
company
division
division
department
... ...
Sept. 2012
Yangjun Chen
ACS-3902
constraint
1:n
1:n
44
Data modeling using the entity-relationship model
Other common situations:
•How would we model bill-of-materials?
•How would we model reporting relationships?
•How would we model geographical relationships?
•How would we model procedures and activities?
•How would we model purchase orders?
•How would we model documents?
… with indexes?
… with versions?
Sept. 2012
Yangjun Chen
ACS-3902
45