8. [Builder] Ontology Data Pipelines24. Writeback Overview

24 - Writeback Overview

This content is also available at learn.palantir.com ↗ and is presented here for accessibility purposes.

One of the key values of Ontology-aware applications is that they permit end-users to make changes to object data. In Foundry, an Action is a transaction that changes the properties of one or more objects. For example, an analyst using a flight alert application may want to assign an alert to a colleague for further work. You'll want that data change propagated back into your Foundry datasets so your colleagues can be sure they're working with the latest information. The process of keeping editable Ontology data in sync with your enterprise data asset is called “writeback.”

This next set of tasks will show you how to configure "writeback" on your three synchronized datasets, but it will not walk through the process of creating actions. For that, follow the Application Developer learning path.

The clickable diagram below illustrates the relevant dynamic whereby ontology edits made through Ontology-aware applications are captured as a dataset back in Foundry. Your colleagues wanting to build analyses and pipelines on top of flight alert data should use Flight Alerts_edited rather than the original backing dataset. This model will take some getting used to, and we will continue to explore it throughout the track.

As a data engineer, you need to know that these "writeback" datasets are updated with changes made through these front-end apps only when their builds are scheduled.