THE DATABASE APPROACH - Database Design and Development: An Essential Guide for IT Professionals - page 26

words, for one occurrence of CUSTOMER, one or more occurrences of ORDER
exist. Similarly, for one occurrence of PAYMENT there are one or more occurrences
of ORDER. By definition, a hierarchical data model cannot represent this kind of
data arrangement with two parents for one child data structure.
Physical Pointers. Just as in the case of the hierarchical data model, related
occurrences of two different record types in a network model are connected by
physical pointers or physical storage addresses embedded within physical records
in the database. Physical pointers link occurrences of an owner record type with the
corresponding occurrences of the member record type. Within each record type
itself the individual occurrences may be linked to one another by means of forward
and backward pointers.
Relational
This topic will provide you with in-depth discussions about the relational model.
This data model is superior to the earlier models. Dr. E. F. Codd, the celebrated
father of the relational model, stipulated the rules and put this model on a solid
mathematical foundation. At this stage, however, we want to introduce the rela-
tional model as a superior data model that addresses the limitations of the earlier
data models.
The earlier hierarchical data model is suitable for data structures that are natu-
rally hierarchical, with each data structure placed at a certain level in the hierarchy.
However, in the business arena, many of the data structures and their relationships
cannot be readily placed in a hierarchical arrangement. The network data model
evolved to dispense with the arbitrary restriction of the hierarchical model. Never-
theless, in both of these models, you need physical pointers to connect related data
occurrences. This is a serious drawback because you have rewrite the physical
addresses in the data records every time you reorganize the data, move the data to
a different storage area, or change over to another storage medium. The relational
model establishes the connections between related data occurrences by means of
logical links implemented through foreign keys. Figure 1-10 illustrates the relational
data model.
Let us note the key features of the relational data model by referring to
Figure 1-10.
Levels. Just like the network data model, no hierarchical levels are present in the
relational model. The lines in a relational data model simply indicate the relation-
ships between the appropriate data structures wherever necessary without the
restriction of connecting only successive levels as in the hierarchical model. As in
the network model, note the lines connecting the various data structures with no
restrictions.
Relations or Tables. The relational model consists of relations. A relation is a
two-dimensional table of data observing relational rules. For example, the CUS-
TOMER relation represents the data content of all customers. The ORDER rela-
tion represents the data content of all orders.