Pipelines

Re-run or roll back a Pipeline

Use Re-run to change Pipeline-wide settings after creation. There is no one-click undo for a finalize, but the destination schema’s versions give you a way back.

In the app: Data Model › Pipelines › your Pipeline › Re-run

Re-run with new settings

Re-run in the mapping editor and Edit pipeline settings on the Pipelines page open the same Update Pipeline dialog. It has the same three steps as the create dialog: Settings, Select Tables, Documentation.

  1. Click Re-run.
  2. In Settings, change what you need:
    • Pipeline Name, Description, and Database System.
    • The naming conventions.
    • Source Schema(s): you can add sources but not remove them. Removing one shows Original source schemas cannot be removed.
    • Destination Schema Name is read-only.
  3. Click Next: Select Tables. Every table and view starts selected. Clear the ones you do not want in the Pipeline.
  4. Click Next: Documentation. Update the documentation if needed. Run AI Rename is already ticked when the Pipeline has documentation; clear it unless you want fresh suggestions, because each run uses AI allowance.
  5. Click Update & Re-run.
[screenshot]
The Update Pipeline dialog on the Settings step, with the read-only destination schema name and the note about source schemas.

What Re-run keeps and what it resets

Re-run rebuilds the mappings from the current sources and your new conventions.

Kept Reset
Column names you renamed yourself or accepted from AI Destination table names go back to what the conventions produce
Pending AI suggestions Each column’s Is PK flag goes back to the source’s primary key
Load strategies and their configuration Every selected table is included again, even if you excluded it in the editor
SK settings and calculated-column choices Column names that came from conventions follow the new conventions

Warning: Re-run overwrites destination table names you edited and key columns you marked by hand. Note them before you re-run, or use Sync from Source if you only need to pick up source changes. See Keep a Pipeline in sync with its source.

Tables you clear in Select Tables are removed from the Pipeline. If one was already finalized, it is marked excluded instead, and the next finalize removes it from the destination schema.

Re-run changes the mappings only. On a finalized Pipeline, validate and finalize again to apply the changes to the destination schema, then regenerate the load code.

Roll back

There is no “undo finalize” button. Your options:

  • Roll back the destination schema. The destination is an ordinary schema with versions. Create a version before you finalize changes, and you can return to it later. See Roll back to a version. Rolling back the schema does not change the Pipeline’s mappings, so a later finalize applies them again.
  • Undo in your database. Generate a migration script between two versions of the destination schema and run it. See Generate a migration script.
  • Start over. Delete the Pipeline and create a new one. The destination schema and its tables stay; delete that schema separately if you no longer need it. See Pipelines list.

Tip: Make Create Version on the destination schema part of your routine before each finalize. It costs nothing and gives you a clean point to return to.