Geography Reference
In-Depth Information
database, each object has a unique identii er (a 'primary key'). If information in other
tables is available, this information can be linked together if both sets of data use the
same identii ers.
With this representation, information is recorded on connectivity (arcs are connected
by nodes), containment (enclosed polygons can be identii ed, although there are none
in the example), and contiguity (arcs shared between polygons, thus determining con-
tiguity, are indicated, e.g. polygons 3 and 4 share arc 5). h e name of the well-known
sot ware environment ArcInfo™ (now part of ArcGIS™) rel ects the relationship between
arcs and attribute information.
2.2.4
Rasters and vectors in GIS software
Most GIS environments allow for conversion from rasters to vectors or from vectors
to rasters. Clearly, the spatial resolution of the raster will limit the positional accuracy
of vector objects generated through raster to vector conversion. For many standard
operations (e.g. measuring proximity to objects; see Section 4.5) either vectors or ras-
ters can be used and so the choice of data model may, for many applications, be less
important than it was in formally less l exible GIS sot ware environments.
Databases
2.3
GIS are ot en dei ned as having at their core a spatial database. In essence, a database
is simply a set of structured information. In order to store, manage, and access com-
puter-based data a variety of database structures have been developed. h e most fre-
quently encountered is based on the relational database structure introduced by Codd
(1970). h e topological structures discussed in the previous section are based on tables
linked by common identii ers, and this is the basis of a relational database. For exam-
ple, in the case of the line topology representation in Figure 2.3, the arcs themselves
and both tables have a unique arc number—there is only one arc with the label '1. h e
information in the two tables can easily be joined together using this common identi-
i er or key. In the case of the polygon topology representation in Figure 2.4, there are
unique arc and polygon numbers. h e polygons listed in the table at the top corres-
pond to the polygons listed in the table at the bottom and the two tables can be linked
using the polygon identii ers. One objective in setting up a relational database is to
store only necessary information and to enable ei cient and ef ective access to the
data. h e setting up of such a database may be non-trivial if there are multiple tables
linked in dif erent ways. While the relational database structure is encountered most
frequently, other database structures are widely used. Object-oriented databases
are quite commonplace; such systems organize data in a structured hierarchy. One
example is the nesting of an individual within a house and a house within a census area
and a census area within a larger administrative unit. Ei cient organization of data
is important for spatial data analysis (particularly in the case of large data sets) as
Search WWH ::
Custom Search