Data editing is necessary for three main reasons:

o Correct input errors.

o Modify data to reflect changes in the real world.

o Enhance the map cartographically.

In the first case, the operator has made the mistakes. In the second case, the data is too old. They both result in databases that are wrong in one or more of the following ways. We will address the third case in the following chapter.

Data with incorrect attributes - Attributes are characteristics of a feature, and wrong attributes would result in the incorrect representation of features. For instance, a piece of property given the wrong owner, the wrong address, or the wrong assessment value is no longer the feature it is supposed to represent.

Data with incorrect position - Spatial data are never one hundred percent accurate because both data collection devices and their operators introduce certain degrees of error. This also causes errors in the coordinate systems because they are established from measurements on the shape of the earth.

Missing data - A highly accurate map may be incorrect because it does not contain all the features intended for the database. On the one hand, some of the highways, houses, and lakes may be missing from a topographic database. On the other hand, the database may contain all the features but miss some of their attributes. For example, it might not contain the assessment value for some of the properties. This might be due to human errors during input, or because such data were not available at the time of input.

Inconsistent data - Logical inconsistency is caused by incorrect relationships among the data items. A common problem is that the positions of geographic features are correct but the relationships among the features are not, causing logical inconsistencies.

In any spatial database, we can expect its data to contain a certain amount of error. For a high quality topographic map, we can only expect 90% of its well-defined features to be within 0.5 mm of their correct positions at the map scale. But an accurate database might not be logically consistent. In the diagram below, the end-points of the two lines are within 0.5 mm of their true positions at map scale and are therefore "accurate" according to the specification for topographic maps. However, the two end-points do not meet as they are supposed to. This inconsistency causes a topological error.

Another example of logical inconsistency is shown in the highway network below. Although the position of the lines are within the accuracy specification, they are not smooth as highways should be. In order words, the position of the points along the line are correct within the specification but are not correct relative to each other.

(288822, 740560), 1:2700

Another common type of logical inconsistency is related to the terrain. In the diagram below, part of the river crosses the same contour three times, giving the impression that it is flowing uphill along part of its course.