Transforms Overview

A Transform node can use either sampled data or all data to calculate statistics. You can use these statistics as a guide to defining one of several transformation.

The following transformations are supported:

• Binning

• Custom

• Missing Values

• Normalization

• Outlier

The transformations available depend on the data type of the attribute. For example, normalization cannot be performed on character data.

You define transformation on a per-column basis. After you have defined a transformation, you can transform several columns in the same way.

The Transform node can run in parallel.

To use a Transform node, connect it to a data flow, that is a Data Source node or some other node such as a filtering node that produces attributes. Then select the attributes to transform.