Sum Attributes

Sum Attributes - Summarize field values of source theme (aggregated spatial join)

With this function you can perform aggregated calculations for several records, i.e. transfer attributes from a source theme to several destination themes by joining the themes spatially (according to the spatial relationship of their features) and by summarizing the field values of all overlapping features (sum, minimum, maximum, majority, standard deviation etc.). This kind of join is only possible between the themes of the active view.

In opposite to the spatial join of ArcView (where only the field value of the first fitting record is used) the field values of all overlapping features are considered with this function. The calculations can be made without priority, weighted by the degree of overlap or reduced proportionally (i.e. the largest area or the longest way inside of the polygon or the nearest situated point have the highest weight). This function is very powerful and of multipurpose use (see Examples of use below) and there are many differences to a normal spatial join (see Hints to join attributes).

You must activate the destination themes, to which you want to sum attributes, before calling the function (click on them with the mouse pointer - keep the Shift key pressed to select multiple themes).

Source Theme: All polygon, line or point themes of the current view are listed. Select the source theme which attributes you want to add to the destination themes.

Field Calculations: In this list each field of the source theme is combined with the possible calculation. You can select several calculations for the same field.
- Majority and Minority: The field value that occurs most frequently or most rarely.
- First and Last: The field value of the record with the smallest or highest record number.
- Count: Number of different values (not the number of records).
- Concat: The field values are concatenated into a text field (e.g. for labeling purposes). The separators between the single values can be defined at Separation String.
- For numeric fields you can also calculate the Sum, Average, Minimum, Maximum, Deviation and Variance.

The order in which these fields are added can be changed with the following buttons:

Move Up : Moves the selected fields one row up. Keep the Ctrl key pressed to move the fields 5 rows up and the Shift key to move the fields to the top of the list.

Move Down : Moves the selected fields one row down. Keep the Ctrl key pressed to move the fields 5 rows down and the Shift key to move the fields to the bottom of the list.

Note: The name of the created fields consists of the first 3 letters of the function name followed by an underscore and the name of the source field (e.g. Min_Area for the minimum value of the field Area, only Minority is abbreviated with Mnr). If the created field names became too long and were truncated, they can be changed with Rename Fields using the alias "A" afterwards.

Calculation Method: There are 3 different ways of weighting the field values according to the degree of overlap between the features (i.e. the intersection between polygons, the length of overlapping lines or the distance to the point):
- Unweighted: The field values are used unweighted and unaltered (i.e. each record is equal, independent from the degree of overlap).
- Weight by Overlap: The field values are weighted by the degree of overlap. This method is suitable only for data that doesn't correlate with area size or line length (e.g. average number of inhabitants per km², per capita income, average age, percentage of forest area, percentage of hardwood, building costs per meter, road width, traffic volume etc.). A value of a feature that is overlapping twice as wide or long, is weighted twice but stays unaltered.
- Reduce by Overlap: The field values are reduced according to the degree of overlap. This method is suitable only for data that depends on the area size or line length (e.g. number of inhabitants per municipality, total turnover, forest area, timber stocks, building costs per road section, passable kilometers etc.). A value of a feature that is only overlapped by 50% is multiplied by 0.5.

Warning: The correct choice of the calculation method is crucial for the correctness or the results. Check your data before each calculation carefully though. If you want to calculate fields that depend on areas and fields that are independent, you have to perform the calculation in two separated steps.

Clear Missing Numeric Fields: Activate this option if the field values of calculations, for which no overlapping features were found, shall be empty (otherwise they get the value 0).

Add Fields for Amount of Overlap: Activate this option to create the fields OvlCnt with the number of overlapping features and OvlSum with the sum of overlapping area, line lengths or point distance.

Additional Numeric Places: You can define the number of additional numeric places for sum calculations. By default the width of these fields is enlarged by 3 places, so up to 999 field values can be added correctly (the probability that more features overlap is relatively small. Because most of those values do not use fully their range, normally even several thousands of values can be added up). If you want to be on the save side, provide more places and use the function Optimize Fields afterwards.

Additional Decimal Places: You can define the number of additional decimal places for the calculation of average, deviation and variance. By default the precision of these fields is increased by 2 decimal places.

Buffer Destination Features: Joining some theme types is only reasonable when the destination theme is buffered (i.e. points or lines are converted to buffer polygons before they are intersected with the source features). Lines can normally only be joined with those points that are situated exactly on the line or with those lines that share an identical line section. You also have to create buffer polygons when joining points to points. In all cases those source features are loaded that are situated within the buffer distance, all features that are more remote are ignored.

Click on

to perform the calculations. Now all features of the source theme are intersected with those of the destination themes and the selected attributes are calculated. This may take some time with large data sets, especially if the features of the destination themes are buffered.

The status bar finally shows the number of calculated attributes and the number of destination records where no overlapping features could be found (and thus no field values could be calculated).

Weighting factor: The field values can be weighted or reduced by the degree of overlap. The calculation of this factor and how the features are related and intersected, depends on the type of the source and destination theme ("buffer" stands for a destination theme of any type that is buffered during calculation):

Source	Destination	Condition for overlap	Weighting factor
Point	Point	coordinate is identically	record number
Point	Line	point is exactly on the line	position along the line
Point	Polygon	point is inside of the polygon	distance to the outline of the polygon
Point	Buffer	point is within the buffer	distance to the outline of the buffer
Line	Point	line contains the point exactly	position along the line
Line	Line	lines are overlapping	length of identical line section
Line	Polygon	line is intersecting the polygon	length of line section within polygon
Line	Buffer	line is intersecting the buffer	length of line section within buffer
Polygon	Point	polygon contains the point	distance to outline of the polygon
Polygon	Line	polygon contains part of the line	length of line section within polygon
Polygon	Polygon	polygons are overlapping	area overlapping with polygon
Polygon	Buffer	polygon overlaps the buffer	area overlapping with buffer
Polygon	=Source	polygons are adjacent	length of the common border line

Adjacent polygons without overlap are normally not considered. If a polygon theme is joined with itself though (destination theme identical to source theme), the adjacent polygons are included and the length of the common border line is used as weighting factor. Polygons that are only connected at a single point are still ignored. Buffer the features to include those neighbors too.

Examples of use:

Calculate for each municipality the most frequent kind of road, the total road length and the average and total building costs.
Calculate for each plot of land the number of adjacent plots, their size and the number of different adjacent municipalities.
Calculate for each forest area the most frequent kind of utilization and the maximum and minimum percentage of hardwood of all other forest areas within a radius of 100 meters.
Calculate for each forest area the number of different landowners and the landowner that owns the majority of the land.
Collect for each forest area the plot numbers in a text field.
Calculate for each river the minimum and mean distance of the nearest houses and collect the house numbers within a radius of 500 meters.
Calculate for each spring the major landuse within a radius of exactly 1000 meters (i.e. weighted by the size of the overlapping area) or the most frequent landuse of all areas situated within a radius of 1000 m (unweighted, i.e. independent from the area size).
Calculate for each house the mean traffic density and minimum distance to the next main road.
Determine for each polygon the major field values of the overlapping points and give higher importance to those points that are closer to its center (i.e. have the greatest distance to the outline).
Determine for each line the major field values of another line theme and give higher importance to those lines which are more similar (i.e. which buffer area has the longest intersecting line segment).
Determine for each polygon the major field values of other overlapping polygons and give higher importance to those that have the most overlap.