DATA MODELING BASICS - Database Design and Development: An Essential Guide for IT Professionals - page 160

Savings Acct
Data View

CustNam CustAddr CustNam CustAddr

CustPhone CustPhone CustName

CustAddr
CustPhone
SavingsAcctNo
SavBalance
InterestRate

CUSTOMER CUSTOMER

PARTIAL
MODELS

CHECKING
ACCOUNT

SAVINGS
ACCOUNT

Checking Acct
Data View

Overdraft

AcctNo
ChkBalance

CustName
CustAddr
CustPhone
ChkAcctNo
ChkBalance
Overdraft

AcctNo
SavBalance

InterestRate

CustAddr
CustPhone CustName

CUSTOMER

CHECKING
ACCOUNT

SAVINGS
ACCOUNT

AcctNo
ChkBalance

Overdraft

AcctNo
SavBalance

InterestRate

FINAL MODEL

Figure 5-16 Integrating partial data models.

3. Ensure correct value sets for each characteristic . If a characteristic of an object
has a defined set of values in one user view and a different set of values in
another user view, this characteristic cannot be taken as the same character-
istic for the object. If the characteristic employee height is given in inches in
one user view and in centimeters in another user view, then there is a dis-
crepancy if the two user views are merged without reconciling the units of
measurement. Review the data model and ensure that the value sets for each
characteristic are correct.
4. Resolve discrepancies in the constraints . Suppose that, in one user view, the
relationship between employee object and department object is such that one
employee may belong to only one department. On the other hand, another
user view such as project assignments may allow one employee to be involved
in multiple departments, that is, assigned to projects of multiple departments.
This is a discrepancy in the constraint on the relationship. Resolve such dis-
crepancies.

One more task remains before the global model is finally complete and ready.
This is the task of validating the data model with the users. Usually, user represen-
tatives knowledgeable in the organization's business processes are selected to
perform the review of the data model. The project team walks the user representa-
tives through the global data model, reviewing each object, each relationship, and
each type of constraint.

Consolidated Data Model

After enhancement and validation, the global data model must be reviewed one last
time to check for any data redundancies. After verification and modification to set