Data Management Techniques in GIS

The ability to efficiently store and quickly access spatial and attribute information is a critical issue for GIS. Database management systems (DBMS) have long been concerned with storage and access issues, but the added complication of managing information in terms of spatial location and proximity has been challenging. There are two fundamental data models used in GIS: raster and vector. They are, in fact, conceptual and organizational approaches to spatial information management that are fundamentally different.

Raster Data Organization in GIS

The raster data model organizes space by tessellation, which is typically a regular square cell (or pixel), though others are possible (e.g., triangle, parallelogram and hexagon). The raster layer is positioned by placing the cell in row 1, column 1. As a result, we require basic information about the cell's size (e.g., 30 × 30 m), orientation (e.g., north), and coordinates. Aside from that, the layer is based on a certain number of rows and columns of cells, so this must be provided as well. A raster representation of an area of interest is defined using this information. The raster structure is notable in that there is no need to store each cell boundary individually because there is spatial regularity. The spatial geometry of the raster model can be easily constructed by knowing the location of the initial cell and other supporting details such as cell size and the number of rows and columns. This enables significant storage efficiencies to be realized.

The raster representation is significant in that we know not only where each cell is located, but also the attribute value associated with each cell. The values of attributes can be a categorical, binary, integer, or real. The GIS must manage the values of each raster cell, which is where coordinated linkage with cell boundaries comes into play. There are several options for an attribute data structure for storing and managing raster data. Some are more cost-effective in terms of storage space than others in terms of access and processing speed.

A Scan order through the raster cells is used to store attribute data for a raster model. A simple scan order is a row by row, row 1, column 1. Attribute data is stored according to the scan order. Row by row, the attribute for row 1, and column 1 cell would be listed first, followed by the attribute for row 1, column 2 cell, row 1, column 3 cell, and so on.

The Morton order differs in the scan of the cells, beginning with row 1, column 1, then row 1, column 2, next to row 2, column 1, followed by row 2, column 2, etc. Alternative scan orders are important because some have greater data compression potential than others, reducing storage requirements. The attributes are then saved in a pairing format, with the first piece of information being an integer count and the second being the actual attribute value for those cells. As an example, the run-length encoded row by row scan order is interpreted as follows: 2 1 indicates that the first two cells have the attribute value 1; 2 3 denotes that the next two cells have the attribute value 3; 4 2 indicates that the next four cells have an attribute value of 2; and so on. Data compression is accomplished by reducing the number of attribute values stored. High compression rates (up to or exceeding a factor of 40) for a raster layer are not uncommon, depending on the scan order and encoding scheme.

Advantages of Raster Model

The raster data model is useful and important because it can accurately represent a study region because an attribute value is assigned to each location. Furthermore, it can do so efficiently by handling cell spatial location and geometry with minimal storage requirements. Other data management approaches for a raster layer, such as quadtrees and other hierarchical data structures, are also possible, offering potential storage and access efficiencies.

Vector Data Organization in GIS

There are several DBMS approaches, including relational, network, hierarchical, object, and object-relational. Relational DBMS has been the dominant data management approach in vector-based GIS, though this has arguably evolved to more of a hybrid object-relational DBMS. Efficient storage and quick access to information are critical for the success of any developed system. In fact, such issues have been responsible for the evolution of DBMS in GIS. In the following section, we will look at object-relational DBMS that can support vector feature spatial representation.

To begin, a relational DBMS organizes data using linked tables. For example, one table could contain information about the attributes of bus stops, another table could contain information about bus stop maintenance, and a third table would contain the positional coordinates (latitude and longitude) of the stops themselves. The need to track vector objects distinguishes object-relational DBMS. On the one hand, the database typically contains a large amount of attribute information, which makes sense to manage using well-developed relational DBMS functionality. Dealing with geographic space and spatial objects, on the other hand, complicates matters. The geometry of vector features varies, which is one of the reasons for this. A point, for example, is always identified by two pieces of information: latitude and longitude. A polyline, however, can be defined by just two points or by hundreds or even thousands of points. According to the line feature. Similar circumstances apply to polygons. Feature geometry and attribute data are now handled in unison by the object-relational DBMS methodology. The object-relational DBMS approach goes beyond storage considerations, though this is covered in more detail in the following section, as access to data and potential operations are its top priorities.

Advantages of Vector Model

• Need a small space or place for storage data (disc).
• One layer can be connected to numerous attributes to conserve storage space for data.
• Makes the link between topology and network very simple
• Display spatial data graphically closely, like a man-handed map;  
• Have a high spatial resolution.
• Easily for making projections and coordinating transformation.
• Very good in correction limits, apparent and clear for making administration maps and owned lands.