Logical Data Modeling
Terms
undefined, object
copy deck
- What is a model?
-
a representation used to help visualize something that cannot be directly observed
- Models are used to (4)
-
Help identify and validate requirements
Organize requirements into coherent pictures
Document and communicate requirements
Support and enhance text-based requirements - Process Models
- Identifies the essential business activities that exist within an organization
- data models
- used to represent the information requirements of an organization
- Entity Relationship Diagram
- used to document understanding of the current or desired information requirements of a business area
- Data Flow Diagram
- used to document understanding of how business processes use data
- Functional Decomposition Diagram
- used to fully describe the activities of a business area
- Business area
- is a naturally cohesive group of activities that share data
-
characteristics of a business area (3)
-
includes all information required by the business
includes all activates performed by the business
can cross organizational boundaries - Perfect Technology
-
Infinite storage capacity
infinite processing speed
correct processing within the business area
availability at zero cost - Assuming perfect technology allows…
-
focus on the analysis, on what the essential business processes are, rather than on how they are or might be implemented
- Processes
- a sequence of activities that has a beginning and an end
- Elemental Processes
- a single specific task performed by the business area
- When to stop functional decomposition (2)
-
1. How the process is accomplished physically rather than what is accomplished
2. The technology or mechanism that is used - Common business processes do the following (4)
-
Carry out exactly the same functional processing
Manipulate exactly the same information
Have exactly the same decomposition structure
Have exactly the same definition - Mini-specification
- the description of an elemental process
- Magical Number 7
-
Optimal number of items that the human brain can remember
Limit each decomposition to between 5 and 9 processes - Techniques to develop a FDD (3)
-
Top Down
Event partitioning (middle-out)
Bottom-up - top-down
-
decompose high level processes to lower levels until elemental processes are identified
Use 7+- 2 rule - Event partitioning
- identify business events and their responses and complete the decomposition from the "middle out"
- Bottom-up
- Start by identifying elemental processes and group up into processes until the top level is reached
- Types on events in Event Partitioning (3)
-
External event
Temporal event
Internal event - External event
- Something initiated by an external agent outside the business area to which the business area must respond
- Temporal event
- Something initiated by the passage of time to which the business area must respond
- Internal event
- Something initiated within the business area to which the business area must respond
- Why use Bottom-Up? (3)
-
You may need to discuss the model at various levels such as executives, management, employees
You may discover missing elemental processes by asking "is there anything else?" at the higher levels
You may want to use the identified proc - Functional Decomposition Diagram can be used to…. (4)
-
Determine the automation boundaries of a project
Analyze process distribution
Provide input to a workflow model
Provide input to a use case scenario - Process Distribution chart shows…
- Where the processes are happening in the organization
- Logical data model
- is a graphical representation of the information requirements of a business area
- The key concept of a logical data model is that it is independent of….
- technology
- Components of a Data Model (4)
-
Entities
Attributes
Relationships
Unique identifiers - Relationships
- show how one entity is associated with another entity
- Unique identifiers
- show how each occurrence of an entity can be separately identified
- A logical data model must (3)
-
contain only business data
use business terminology
assume perfect technology - a Logical data model must not (3)
-
Reflect physical implementation or constraints
Include technology- dependent data (foreign keys, switches, flags)
include derived or redundant data - Derived data is not essential because…
-
it can be calculated or summarized from the values of other attributes in the data model
- 3 levels of data modeling
-
Conceptual
Logical
Physical - Conceptual data model
- high-level logical data model that focuses on how the business world sees data
- Physical Data Model
- Describes the implementation of a logical data in a specific database management system
- Entities represent…
- a group of uniquely identifiable people, places, things or concepts about which the business area needs information
- Naming Standards for entities
-
the name is usually a noun or noun phrase
Usually entered in capitals
Always singular
Uses business terminology - 4 questions to figure out if it is an entity
-
Does the entity have business significance to this business area?
Does the entity have characteristics that this business area needs to use?
Can each occurrence of the entity be uniquely identified?
Does the entity have any relationship to - Attributes
- are the pieces of data the business area needs about each entity
-
Unique identifiers
- show how each occurrence of an entity can be separately identified
- Attribute is…
- a piece of information that describes an entity, it is a characteristic of the entity
- Is it an attribute or not? (4 questions)
-
1. Does it describe the entity further?
2. Does it have significance to this business area?
3. Does it represent an atomic piece of data?
4. Does it apply only to this particular entity? - Analysing Attributes Questions (3)
-
Is it mandatory or optional?
Is it multivalued or single-valued?
Is it unique or not? - Unique identifiers made up of more than one attribute are called?
- Concatenated unique identifiers
- Surrogate unique identifiers are assigned when… (2)
-
There are no attributes that meet the requirements for a unique identifier
There is some reason why a candidate identifier should not be used - Super type / sub type
- a subtype entity implicity inherits all the atributes and relationships of the entity of its super type
- 2 reasons to make a subtype
-
1. entity has many attributes that are not applicable to all occureences of the entity.
2. An entity has many relationships that are not applicable to all occurences of the entity
-
subtype discriminators
-
is an attribute that allows an occurrence of the super type to be classified as an occurrence of a subtype
-
attributive entities
-
is an entity that depends on another entity for its existence
-
2 reasons to make an attributive entity
-
1. if a attribute is multivalued
2. if there are attributes that are not dependent on the unique identifier.
-
Attributive entity naming convention (2)
-
1. use business name
2. use the name of the original entity + name of moved attribute
- attribuutive entity unique identifier is usually a
- concatenated ID
- an attributives orignal entity always has what relation on it
-
| |
-
associative entities
-
an entity that describes a relationship or association between other entities
-
Reason for creating an associative entity (1)
- there was a MANY TO MANY relationship
- associative entity naming standard
-
if no business name try combining the names of the 2 original entities.
- Associative entities unique identifier (2)
-
Try combining the UIDs of the original entities.
If not sufficient add an attribute.
-
Associative entities relationships usually become
- 1 and only 1 to x and many
-
data constraint
-
identifies the type of data that may be used to represent values of the attribute
- Data flow diagrams identify
-
the processes of a business area that transform data
-
Context level data flow diagram contains only...
-
the highest level process and the external agents with which it communicates data
- external agents
- a perso, organization, or system outside the business area
- data flow is also called a
-
pipeline
- Packet
-
data that flows through the pipeline
-
data store
-
data at rest
-
leveled data flow diagrams are produced by
-
hierarchical decomposition from the context level diagram