Overview of Nodes

When creating an extension that implements a new node, you need to become familiar with the characteristics of IBM® SPSS® Modeler nodes. Doing so will help you to define them correctly in the specification file.

IBM SPSS Modeler nodes are classified into source, process, output, and modeling nodes, depending on their function. In CLEF, nodes are classified in a slightly different way. The mapping between the two systems is shown in the following table.

Table 1. CLEF node types
IBM SPSS Modeler classification	Palette	CLEF node type
Source nodes	Sources	Data reader
Process nodes	Record Ops	Data transformer
	Field Ops
Output nodes	Graphs	Document builder
	Output (Report node)
	Export	Data writer
Modeling nodes	Modeling	Model builder

When you create a new CLEF node, you define it as being one of the CLEF node types. The node type you choose depends on the main function of the node.

Table 2. Node types and functions
CLEF node type	Description	Corresponding node palette	Icon shape
Data reader	Imports data into IBM SPSS Modeler from a different format.	Sources	Figure 1. Source node shape (circle)
Data transformer	Takes data from IBM SPSS Modeler, modifies the data in some way, and returns the modified data to the IBM SPSS Modeler stream.	Record Ops, Field Ops	Figure 2. Ops node shape (hexagon)
Model builder	Generates models from data in IBM SPSS Modeler.	Modeling	Figure 3. Model builder node shape
Document builder	Generates a graph or report from data in IBM SPSS Modeler.	Graphs	Figure 4. Graphs node shape (triangle)
		Output (Report node)	Figure 5. Output node shape (rectangle)
Model applier (also known as a "model nugget")	Defines a container for a generated model that is brought back to the IBM SPSS Modeler canvas.	–	Figure 6. Model applier node shape (gold diamond)
Data writer	Exports data from IBM SPSS Modeler format to a format suitable for use with another application.	Export	Figure 7. Export node shape (rectangle)

You define the node type, together with other attributes, in a Node element in the specification file—for example:

<Node name="sort_process" type="dataTransformer"
      palette="recordOp" ... >
     -- node elements --
</Node>

The palette attribute defines the palette in the IBM SPSS Modeler main window from which users will be able to access the node--in this case, the Record Ops palette. If you omit this attribute, the node appears on the Field Ops palette.

A number of example nodes are supplied with IBM SPSS Modeler. See the topic About the Examples for more information.