Logical Database Design Nazife Dimililer. II - Logical Database Design Two stages –Building and...

Logical Database Design

Nazife Dimililer

II - Logical Database Design

• Two stages – Building and validating local logical model

– Building and validating global logical model

The local model represents different (or individual) user viewsWhat payroll sees vs. what the accounting sees

The global model represents ALL user views What will be the requirements of all departments combined?

Stage 1: Local logical model(Step 2)

(for relational model)

1. Remove features not compatible with the model you choose (relational model)

2. Derive relations for local logical model3. Validate relations using normalization4. Validate relations against user

transactions5. Define integrity constraints6. Review local logical data model with the

user

Stage 2: Build and validate global logical data model

(Step 3)

1. Merge local logical data models into global model

2. Validate global logical data model

3. Check for future growth

4. Review global logical data model with the users

2.1. Remove features not compatible with the relational model

Objective: Refine the local conceptual data model to remove undesirable/unsupported features to map to local logical model

i. Remove M:N Relationshipsii. Remove complex Relationshipsiii. Remove relationships with attributesiv. Remove multivalued Attributesv. Remove complex Attributesvi. Remove redundant relationshipsvii.Remove inheritance relationshipsviii.Re-examine 1:1 relationships

2.1.i. Remove M:N Relationships

Exam

ple 1

hasreferToregisterFor Take

Grade

Course

CourseCodeTitle

CourseGroup

GroupNo

Student

StudentIdName

2.1.i. Remove M:N Relationships

Exam

ple 1

Solutio

n

2.1.i. Remove M:N Relationships

hashasProduct

ProductIdNamePrice

BuyCustomer

CustomerIdName

Solutio

n?

Exam

ple 2

Solutio

n?

Solutio

n 1

2.1.i. Remove M:N Relationships

PK of the invoiceline entity is {invoiceId,ProductId}. This means that the same product cannot appear twice on an invoice! In supermarkets, when you buy two of the same product, each of them may be ringed separately. Therefore this solution is not suitable for supermarkets.

2.1.i. Remove M:N Relationships

gets

has

containsInvoice

Invoice|dInvoiceDate

InvoiceLine

LineNoQuantity

Product

ProductIdNamePrice

Customer

CustomerIdName

Solutio

n 2

Another problem with the previous solution is that, if you need to print (select from) the contents of the invoiceline entity, you may get a different order every time. But when it comes to invoices, the order is important and every print-out must look exactly identical. For this reason and to solve the problem mentioned before, it is customary to add lineno attribute to invoiceline entity. Please notice that the invoice line entity will not depend on the product entity in this case.

2.1.ii Remove complex Relationships

Exam

ple 1

2.1.ii Remove complex Relationships

Solutio

n?

MadeBy

SignedByMadeFor

Contract

ContractDateAmount

Movie

MovieIdNameBudget

Studio

StudioIdName

Star

StarIdName

2.1.ii Remove complex Relationships

Solutio

n

MadeBy

SignedByMadeFor

Contract

ContractIdContractDateAmount

Movie

MovieIdNameBudget

Studio

StudioIdName

Star

StarIdName

2.1.iii Remove Relationships with Attributes

• Relationships with Attributes are usually M:N or complex relationships. We dealt with these in step 2.1.i and step 2.1.ii

• If a M:1 relationship contains an attribute move the attribute in the same direction as the Foreign Key.

• If a 1:1 relationship has an attribute move it to either of the participating entities

2.1.iii. Remove Relationships with Attributes

Exam

ple 1

Solutio

n

Work

Department

DepartmentIdName

Employee

EmployeeIdNameStartDateEndDate

Foreign Key Moves in This Direction

Is the relationship really

1:M?

2.1.iii. Remove Relationships with Attributes

Exam

ple

Solutio

n

Manage

Department

DepartmentIdName

Employee

EmployeeIdNameStartDateEndDate

Manage

Department

DepartmentIdNameStartDateEndDate

Employee

EmployeeIdName

OR

Is the relationshi

p really 1:1?

has Phone

PhoneNo

Student

StudentIdNameMotherNameFatherName

2.1.iv. Remove Multivalued Attributes

Exam

ple

Solutio

n

Assumptions:

Each student has only two parents: mother and father

Store as many phone numbers for each student as needed!

2.1.iv. Remove Multi-valued Attributes

belongsTo

PhoneType

TypeCodeDescription

has

Phone

PhoneNoDescriptionStartingHourEndHour

Student

StudentIdNameMotherNameFatherName

belongsTo

PhoneType

TypeCodeDescription

has

Phone

PhoneNoDescriptionStartingHourEndHour

Student

StudentIdNameMotherNameFatherName

Enhancing the previous solution : Add new attributes/entities/relationships

Another solution : Assume that two students may use the same phone number

2.1.v. Remove Complex Attributes

Exam

ple

Solutio

nStudent

StudentIdFirstNameLastNamePhoneNo

If the complex attribute may be shared by one or more

entities consider

representing it as a separate

entity

2.1.vi. Remove redundant relationships

Exam

ple

Solutio

n

givesgets

Salesperson

IdName

Invoice

InvoiceNoInvoiceDate

Customer

CustomerIdName

2.2. Derive relationships from the logical data model

Objective : derive tables to represent entities, relationships, and attributes defined in the user view. The local logical model at this step includes only the following

• Strong entities• Weak entities• Strong 1:M Binary relationships• Strong 1:1 Binary Relationships• Identifying (Weak) relationships

Document Relations and Foreign Key attributes

2.2 Validate Model Using Normalization

• Normalization is a procedure for deciding which attributes belong together in an entity type.

• Normalization forces us to understand each attribute represented in the database completely.

• Normalization minimizes redundancy and thus update and deletion anomalies.

• Derive functional dependencies and normalize all relations up to BCNF!

There are arguments

against normalizatio

n

Some times you need to

Denormalize

Denormalization

• Transforming normalized relations into unnormalized physical record specifications

• Benefits:– Can improve performance (speed) be reducing number of table

lookups (i.e reduce number of necessary join queries)

• Costs (due to data duplication)– Wasted storage space– Data integrity/consistency threats

• Common denormalization opportunities– One-to-one relationship– Many-to-many relationship with attributes– Reference data (1:N relationship where 1-side has data not used

in any other relationship)

2.4 Validate against user transactions

Objective: To ensure that the local logical data model supports the transactions required by the user view

There are two approaches:• Check that all information (entity, relationship,

attribute) required by the transaction is provided by the model by documenting a verbal description of the transaction

• Represent the pathway taken by the transaction diagrammatically on the ERD

2.5 Draw Entity Relationship Diagram

• Draw a final ERD that is the local logical representation of the data given in a user’s view of the enterprise.

2.6 Define Integrity Constraints

• Integrity constraints are the constraints that we need to impose in order to protect the database from becoming inconsistent.

• The following constraints must be included in your design:– Required Data– Attribute Domain Constraints– Entity Integrity– Referential Integrity– Enterprise Consttraints

2.6 Define Integrity constraints II

• We have dealt with the required data, attribute domain and entity integrity constraints before. Revise the design to make sure they are correct and complete

• Referential Integrity constraint must be further qualified to include existence constraint.– Existence constraint defines conditions under

which a unique key or foreign key may be inserted, deleted or updated.

2.6 Define Integrity constraints III

A foreign key definition should include one of the following :

• No action : Do not allow • Cascade : Cascade the changes to the children• Set Null : set the value of the children to null• Set Default : Set the value of the children to a

default value• No Check : Do not heck any thing (This is not a

desirable option)

2.6 Define Integrity constraints IV

• Enterprise constraints are also known as business rules.– Constrained by enterprise rules governing

the “real world” transactions– Usually include more than one entity or a

complex rule based on aggregate values.

• Document of all the integrity constraints. This an important step in the logical design process!!

2.7 Review local logical data model with user.

Objective: To ensure that the local logical data model is a true representation of the user’s view.

Stage 2: Build and Validate Global Logical Data Model

(Step 3)Objective: Combine individual local logical

data models into a single global logical data model representing the part of the enterprise we are interested in modeling

This is stage 2 in the Logical design and Step 3 of database design.

3.1 Merge local logical data models into global model

Objective : Merge local logical data models into global model Some typical tasks

– Review the names and contents of entities/relations and their candidate keys

– Review the names and contents of relationships and foreign keys – Merge entities (relations) from the local models – Include, without merging, entities (relations) unique to each local

data model– Merge relationships (foreign keys) form the local data models– Include (without merging) relationships (foreign keys)– Check for missing entities (relations) and relationships (foreign

keys)– Check foreign keys– Check integrity constraints– Draw global ER diagram– Update documentation

3.2. Validate global logical data model Validate global logical

data model

• Validate the global logical data model using normalization and against the required transactions.– Carefully inspect the boundaries!

3.3 Check for future growth

Objective : To determine whether there any significant changes likely in the foreseeable future and to assess whether the global logical data model can accommodate these changes?

Develop a model• that is extensible, • that can evolve with minimal effort • That can evolve with minimal effect on

existing users

3.4 Draw Final Entity Relationship Diagram

After the validation we have a final ERD representing the data requirements of the part of enterprise to be modeled.

All the other documentation should also be updated to reflect any changes.

4. Review global logical data model with users

Objective: To ensure that the global logical data model is a true representation of the entreprise.

Final Word

• Remember that the goal of the DB development is to produce a DB that provides an important resource for an organization.

• The DB should be designed so that it can serve the customers and other team members efficiently.