Database Reference
In-Depth Information
of creating data models. We went through the methods and steps for creating the
object-based data model and the entity-relationship data model. Now the task is to
create a relational data model, which is not the same as one of those two data
models. Why do you need to create a relational data model? If you are developing
a relational database system, then you require that your information requirements
be represented in a relational data model first. Let us explore the methods for cre-
ating a relational data model.
Requirements for Data Model
From the previous chapters you know how to create a semantic data model from
the information requirements. A semantic data model captures all the meanings and
content of the information requirements of an organization. We also discussed the
merit of the semantic data model. Being a generic data model, it is not restricted
by the structure and format rules of the conventional data models such as hierar-
chical, network, or relational data models. You are now convinced that repre-
senting information requirements in the form of a semantic data model is the
proper way.
Well, what are the steps between creating a semantic data model and the imple-
mentation of a relational database system for your organization? You know that the
semantic data model, if created correctly, will represent every aspect of the infor-
mation that needs to be found in the proposed database system. The next steps
depend on the extent and complexity of your database system. Let us examine the
options.
Design Approaches
Database practitioners adopt one of two approaches to design and put together a
relational data model. The relational data model must, of course, truly represent the
information requirements. In the simplest terms, what is a relational data model? It
is a collection of two-dimensional tables with rows and columns and with relation-
ships expressed within the tables themselves through foreign keys. So, in effect,
designing and creating a relational data model reduces to creating the proper col-
lection of two-dimensional tables.
Figure 8-19 presents the two design approaches for creating a relational data
model.
Note how in one approach, you go through the steps of creating a semantic data
model first and then transform the semantic model into a relational data model. The
other approach appears to be a short-cut method bypassing the semantic data
model. In this approach, you proceed to the task of creating the relational data
model straight from the requirements definitions. Let us examine the basics of the
two approaches.
Semantic to Relational Model
The first method shown in Figure 8-19 takes you through the semantic data model.
In this approach, you complete the semantic data model. To create the semantic
data model, you may use either the object-based data modeling technique or the
Search WWH ::
Custom Search