Page tree

Trifacta Dataprep



Contents:

If you licensed Dataprep by Trifacta before Oct. 14, 2020, you are using the Dataprep by Trifacta Legacy product edition. On October 14, 2022, this product edition will be decommissioned by Google and will be no longer available for use. Current customers of this product edition are encouraged to transition to one of the product editions hosted by Trifacta. See Product Editions.

   

Feature Availability: This feature is not available in
Dataprep by Trifacta Legacy only.
Trifacta Photon is an in-memory running environment that is hosted on the Trifacta node. This environment is initialized only when a job is queued for execution on it. Designed for small- to medium-sized jobs, it offers superior performance due to its location on the Trifacta node and its in-memory processing.

When you choose to run a job in the Trifacta applicationTrifacta Photon is selected as the default running environment if it is available and the job size is estimated to small or medium.

Tip: Trifacta Photon is also used for sampling jobs that are configured to use the Quick Scan method. For more information, see Overview of Sampling.


Tip: In the Run Job page, select Photon to run the job on this running environment.

Trifacta Photon is enabled by default but can be disabled as needed.


NOTE: Trifacta Photon may require enablement in your project of workspace. See Configure Running Environments.

NOTE: Jobs that are executed on Trifacta Photon may be limited to run for a maximum of 10 minutes, after which they fail with a timeout error. If your job fails due to this limit, please switch to running the job on Dataflow.

NOTE: Trifacta Photon cannot process numeric values with more than 16 digits. Columns containing such values are converted to String values, and the digits beyond 16 are converted to 0.

NOTE: When a recipe containing a user-defined function is applied to text data, any null characters cause records to be truncated during Trifacta Photon job execution. In these cases, please execute the job on Spark.

This page has no comments.