Result Write-back

Result Write-back

Map workflow result paths to columns in the workflow:

{
  "data": {
    "source": "users",
    "results": {
      "created_user_id": "data.userId",
      "confirmation": "data.confirmationText"
    }
  }
}

The runner also maintains reserved columns: __automation_status, __automation_run_id, __automation_started_at, __automation_completed_at, __automation_duration_ms, __automation_attempts, __automation_result, __automation_error_type, and __automation_error.

Missing result columns are added to the header. Updates are sparse, so workflow write-back does not overwrite source cells.

data.results values are nested paths into the row’s FlowExecutionResult; extracted values normally live under data.<name>. Failed rows populate the error columns, while successful rows store the extracted result JSON in __automation_result.