Choosing a preparation run method
Last updated: 10/30/2026
Choose how to run a preparation based on its destination, automation needs, and execution environment.
You can run a preparation from the Talend Data Preparation interface, automate runs with the Talend Data Preparation Run API, or apply a preparation in a Talend Studio Job with the tDataprepRun component.
| Method | Use it to | Requirements and constraints |
|---|---|---|
| Run from the interface | Send prepared data to a new or existing dataset, update the source dataset, or download the result. | For Talend Cloud runs, use Remote Engine Gen2 R2024-06 or later, or Cloud Engine for Design. Restart the Engine to enable the latest updates. The interface runs only the current preparation version. For steps and destination limitations, see Running a preparation. |
| Automate with the Talend Data Preparation Run API | Schedule or integrate preparation runs by calling the API to list preparations, start a run, monitor its status, and retrieve run history. | You must have access to the preparation and destination, and the preparation must have completed a successful manual run. Direct download is not supported as a destination. API run history is separate from the interface run history. For details, see Automating a preparation run via API. |
| Run in a Talend Studio Job | Apply a preparation in a data integration or Big Data Job with the tDataprepRun component. | A data integration Job processes the preparation on the Talend Data Preparation server. A Big Data Job retrieves the preparation definition from the server and processes it on the cluster. Use a preparation version to keep the Job's result predictable. See tDataprepRun workflow in a Talend Job and Using a version in a Talend Job. |
Information noteNote:
When you automate runs with the Talend Data Preparation Run API, its run configuration comes from the latest successful manual run. Use a dedicated preparation for automation if later manual runs could change that configuration.
For a database destination, updating more than 100,000 rows in an interface run can cause performance issues. For more information about creating and managing preparations, see Adding a preparation.