Create Custom Transformer

A Create Custom Transformer is an operation used to create a user-defined Transformer.

The main purpose of creating a custom Transformer is to transform multiple DataFrames using the same logic.

A Custom transformer can be created by building an internal workflow from operations. The internal workflow must have two following nodes:

• input node (“source”) which provides an input DataFrame,
• output node (“sink”) which receives the result of executing the transformation on the input DataFrame.

The Create Custom Transformer operation enables the user to create a transformation by defining a data flow between the input and output nodes. For the internal workflow to be valid, the same requirements as for a regular workflow must be met, i.e.:

• it must not contain a cycle,
• the parameters of all operations must be correct,
• each input port must have exactly one incoming connection,
• the classes of entities passed between ports must meet requirements of these ports’ type qualifiers.

In addition to that:

• it should take the input DataFrame from the input node,
• it should output the transformed DataFrame to the output node.

The Transformer produced by the operation can be used as an input to e.g. a Transform operation. When the Transform operation is executed, its input DataFrame is passed to the input node of the internal workflow of the custom transformer. Once the internal workflow is finished, the result of the output node is returned as the output DataFrame of the Transform operation.

Since: Seahorse 1.0.0

Input

The Create Custom Transformer does not take any input.

Output