developing-datacloud-code-extension

To build and deploy Data Cloud Python transformations, you scaffold a project using the SF CLI data code extension plugin, develop batch or real-time logic, scan for permissions, test locally, and deploy via a payload directory.

The scan process generates the necessary permissions and configuration files, such as config.json, to validate what Data Cloud DLO and DMO access the extension requires before you run or deploy it.

Python 3.11 and the Data Cloud custom code SDK are required to develop custom code, while Docker is optional and specifically needed for the deploy workflow to package the extension correctly.

Script-based transformations handle batch processing, whereas function-based transformations process real-time data, both reading from DLOs or DMOs and writing transformed results back to Data Cloud.

You test Data Cloud custom code locally by running the extension against an authenticated Salesforce org to verify the transformation logic and permissions before executing the deployment workflow.

Data Cloud code extension deployments require strict packaging via the payload directory layout, and failing to structure your project files correctly in this directory will cause the deployment to fail.