Edit Transform Node

You can define and edit the Transform node using the Edit Transform Node dialog box. This dialog box has two tabs:

• Transformations

• Statistics

In the Transformation tab, the statistics for each column is displayed. If you do not want to see statistics, then deselect Show Statistics.

You can perform the following tasks in the Transformations tab:

• Define Transformation: Select one or more original columns, that is, columns that are not transformed. Click .

  The Add Transform dialog box opens if you selected one or fewer columns. Otherwise, the Apply Transform Wizard opens.

• Define Custom Transformation: Select one or more original columns, that is, columns that are not transformed. Click .

  The Add Custom Transform dialog box opens. You can add custom transformations here.

  The default behavior is to ignore the original column and to use the transformed column as output. The values displayed in the Output column are indicated by:

  • : For a column that is included

  • : For a column that is ignored

• Change Values in the Output Column: Click the icon displayed in the Output column to edit the values in the Add Transform dialog box.

• Edit Transformed Column: You can edit transformed columns only. For example, you can edit AGE_BIN, but not AGE. To edit transforms, select one or more transformed columns and click . The Edit Transform dialog box opens if you selected one or fewer columns.

• Delete Transform: Select one or more transformed columns and click .

• Filter Column: To limit the number of columns displayed, click . You can search by:

  • Output Column

  • Transform

  • Source Column

• Clear filter definition: To clear filter definitions, click .

• View effects of transform: To view the effect of a transformation:

  • Run the node.

  • After running of the node is complete, double-click the node.

  • Select the transformed column to see histograms that compare the original column with the transformed column.

When a column has a transformation applied to it, a new row in the list of columns is generated. Because each column must have a name, the name of the new row is based on the name of the old column and the type of transformation performed. Typically users want to transform a column and only have the output of the transform node contained in the new column. The original column has an option set to prevent it from being passed as one of the output columns. For example, if you bin AGE creating AGE_BIN, then AGE is not passed on and AGE_BIN is passed on.