Custom visual transform

You can create your own visual transforms and use them in Visual ETL jobs in Amazon SageMaker Unified Studio. To do this, you create a JSON config file and a Python file with your desired custom transform code and settings. Then you upload both files to the transforms folder in your project's Amazon S3 shared folder. The transforms then appear in the Visual ETL transform library.

Migration from Glue Visual ETL

If you already created custom visual transforms in AWS Glue under the same account, those transforms would also be available in Amazon SageMaker Unified Studio Visual ETL. For more information about custom visual transforms in AWS Glue, see Transform data with custom visual transforms in the AWS Glue Developer Guide.

The following sections describe the JSON config file schema, the Python file, and the steps to create and use a custom visual transform.