HANDLING INTEGRITY CONSTRAINTS OF COMPLEX OBJECTS 
IN SPATIAL DATABASES 
Hanna H. Kemppainen 
Researcher 
Finnish Geodetic Institute, Finland 
Commission III 
ABSTRACT 
The study reported in this paper is based on two assumptions about the importance of the consistency of a 
geographical database. Firstly, consistency is an important factor concerning the quality of data. Secondly, 
consistency of a geographical object requires that the description of the context of the object is a part of the object 
definition. 
In object-oriented databases, it is possible to model geographical entities as complex object types whose semantics 
are captured through constraints on objects and object relationships.This paper examines the modelling of 
geographical objects and spatial relationships between them. A geometrical object model for geographical entities 
is presented which consists of 1) the structural definition of the objects, and 2) the enforcement of the implicit 
structural constraints as object methods. A modelling technique for specifying explicit constraints on spatial 
relationships between objects is also presented. 
KEY WORDS: Geographical modelling, Integrity constraints, Object-oriented databases 
1INTRODUCTION 
1.1 Motivation 
Consistency of a geographical database is an 
important factor of quality of geographical data. The 
description of quality of data provides information 
for a user to evaluate the fitness of the data for a 
particular use. This study is motivated by the lack of 
support for user-defined constraints in commercial 
GIS software. As to the author's knowledge, there are 
no such systems where the user might specify 
arbitrary constraints on the database. The purpose of 
this paper is to present issues related to the 
consistency of the data in the database, what is the 
meaning of constraints of the database and what the 
constraints on geographical database might be. These 
issues are handled in order to show what could be 
required from a software system that manages a 
geographical database as far as the consistency of the 
data is concerned. 
The specification and enforcement of integrity 
constraints in object-oriented databases is studied also 
in this study. This is done to show the possibilities of 
high level data models to capture semantics of data, 
which capability is lacking from conventional, record 
based data models. The significance of high level data 
modelling lies in the understandability of the 
concepts to the end-user of the system, and hopefully 
also as a programming work simplifying agent. 
A database is a representation of some portion of the 
real world, according to the requirements of the 
application that the database is going to support with 
its data. The requirements for data are specified in the 
database design phase, and they are formalized in the 
database schema. The database schema is an instance of 
some data model. The database schema is the 
definition of the correct, allowable types of data, whose 
instances may be stored in a particular database. The 
schema thus constrains the instances of data types to 
some pre-defined structure. The requirements that 
cannot be presented in the schema are presented 
outside the schema, for example, in application 
programs. The contents of the database is constrained 
in much the same way by these two types of 
constraints; only their specification is different. 
Consistency is probably the most important criterion 
for the use of the database once its usability has 
otherwise been demonstrated. As a database serves as 
an information-providing element in an information 
system, the user of the database has to be sure that he 
can trust the image of the real world the database is 
offering. Therefore, the user of the database has to 
know the semantics of the constraints put on the 
database and the processes by which the constraints are 
enforced; he also has to understand the results of the 
constraint enforcement. 
An example of an integrity constraint referring 
specifically to a geographical database is that a location 
occupied by an object representing a house may not be 
occupied by an object representing a lake. Integrity 
enforcement automatically relieves the user from 
checking the database by hand, or rather, by eye, that 
every lake in the database is free from houses. The 
456 
spe 
dep 
det: 
are 
121 
The 
exal 
con: 
rep: 
thu: 
dat: 
con: 
are 
cort 
em} 
whi 
data 
In tl 
of 
geo 
and 
moc 
geo 
stru 
geo 
the 
Adc 
defi 
data 
The 
dat: 
thir 
met 
kno 
Sarj 
orie 
app 
infc 
In 
inte 
disc 
geo 
Seci 
des 
enti 
MO 
a st 
app 
and 
13: 
The 
and 
geo 
con 
thes 
in t 
Geo 
abo