Skip to main content
The live transformation surface has two main parts:
  • a branch-scoped Python script in the Transform panel
  • a separate Filter node for simple keep-or-drop rules
There is no standalone visual rename, cast, or visual-mapping workflow in the supported product path today. Those outcomes are implemented in Python.

What is live today

  • branch labels in the Transform panel
  • source table or resource selection from inside that panel
  • a Python editor that expects def transform(record):
  • a Load Example menu when example packs exist for the current source type
  • Preview with sample data before saving
  • a dedicated Save Script action
  • a Filter node with Python Expression and SQL WHERE tabs

What is not the supported workflow

  • a point-and-click rename screen
  • a separate cast wizard
  • a visual fill-nulls screen
  • standalone visual-mapping documentation as the primary production path
  • a separate transform administration workflow
  • a user-facing toggle for skipping transform errors; saved scripts currently use fail-on-error behavior

Fast path

  1. Select the source table or resource.
  2. Keep the default pass-through script first.
  3. Preview the current payload.
  4. Add one transformation step.
  5. Preview again.
  6. Save the script.
  7. Validate the destination preview and then run manually.