Before tackling this tutorial, you will need to download and install a dataset following these instructions:
Create a folder called multipart somewhere under your personal directory (e.g. C:\Users\jdoe\Documents\Tutorials\multipart\).
Download the data for this exercise then extract the contents of multipart.zip into your newly created multipart folder.
Open the multipart.aprx file.
The project consists of a Counties layer and a Parks layer.
Right-click on the Counties layer in the Contents pane and select Attribute Table.
This will bring up the layer's attribute table.
The State of Maine is made up of 16 counties. Each county is assigned its own record in the table.
In the attribute table, select Cumberland county. This should highlight the county in the map.
Zooming in on Cumberland County, you'll note that the feature is made up of multiple polygons--each linked to the same record in the attribute table. We say that we have a many-to-one relationship between the geometric elements (polygons in this example) and the attribute records. This is an example of a multipart feature dataset.
A many-to-one relationship has its benefits in that you do not need to unnecessarily replicate the same records multiple times for a group of features sharing the exact same values. This helps reduce the attribute table size. In our working example, the Counties layer has more than 6233 polygons, yet we only need 16 rows of attribute values .
There are, however, scenarios where a one-to-one relationship between geometric elements and attribute records may be desirable. For example, you would not be able to compute the surface area for each individual polygons since only a single common value could be computed for all polygons associated with a same county. Another example where a one-to-one relationship may be needed is in selecting individual features. Note that selecting one polygon in the Counties layer will select all polygons sharing the same attribute record.
To create a one-to-one relationship from a many-to-one relationship, we need to convert the GIS layer to a singlepart feature. This step is shown next.
Before running the geoprocess, make sure to clear any selections you might have in the Counties layer by clicking in the Clear button under the Map tab. If the Clear button is ghosted out, you do not have a selection in play.
If you run a geoprocess when a selection is active in the input layer, the geoprocess will only apply to selected feature(s).
In the Analysis ribbon, click on the Tools button.
In the Geoprocess pane, type "singlepart" in the search bar.
This should bring up the Multipart to Singlepart tool. Click on that link to bring up its pane.
Select the Counties layer in the Input Features field.
Name the output shapefile County_multi.shp. Be sure to save it in the multipart project folder.
Click Run.
The output shapefile should look identical to the input layer. However, if you open its Attribute table, you should see 6233 records--one for each individual polygon.
Note that the attribute fields carry over from the original input layer. A new field is also added, ORIG_FID, which references the FID from the input layer. ArcGIS simply replicates values from the input layer as needed.
When replicate values are not need, you might want to consolidate the features so as to create a many-to-one relationship such as the one we encountered with the Counties layer. This can be computationally beneficial by reducing the size of your attribute table. In this next example, you will convert the Parks layer (which is a singlepart feature layer) to a multipart feature layer using the dissolve tool.
Open the Parks layer attribute table.
Sort the Name column in ascending order by right-clicking the NAME field and selecting Sort Ascending.
You'll note the many instances of the same park name. We will consolidate the records such that each park name appears just once in the attribute table. This will require that we dissolve the layer by a common park name (the Name field in this example). Note, however, that the Acres column does not share the same unique value across the same park name. This will pose a problem since all the different acreage values will need to be consolided to a single value. We can either chose to ignore this field and not have it present in the dissolved layer, or we can chose a summary statistic to reduce the value to a single one for each unique park name. Here, we'll summarize the acreage value to come up with a single overall value for each park.
In the Geoprocess pane search bar, type Dissolve.
Click on the Dissolve tool.
Set the input field to Parks.
Name the output shapefile Parks_multipart.shp.
Set Name as the dissolve field.
Under the Statistics Fields section, select the Acres column and set the summary statistic to Sum.
Click Run.
Note that we went from 832 records to 172 records such that each record in the parks_multipart shapefile has a many-to-one relationship between the attribute records and the geometric elements (polygons in this example).
This wraps up this tutorial.