Study Databricks DE-PRO Delta Sharing and Federation: key concepts, common traps, and exam decision cues.
The quickest DE-PRO mistake here is collapsing three separate ideas into one: share data to another Databricks deployment, share to an external platform, or query an external system through federation.
| Requirement | Better first instinct |
|---|---|
| share live governed data to another Databricks deployment | Databricks-to-Databricks sharing |
| share governed data to a non-Databricks platform | open sharing protocol |
| query external systems through Databricks governance | Lakehouse Federation |
| If the data should… | Stronger first answer |
|---|---|
| stay in Databricks and be published out | Delta Sharing |
| stay in an external system and be queried through Databricks governance | Lakehouse Federation |
| be shared specifically to another Databricks deployment | Databricks-to-Databricks sharing |
This is the fastest way to separate sharing from federation.
| If the stem says… | Strong reading |
|---|---|
| “share to another Databricks workspace” | stay in the Databricks sharing lane |
| “share to any computing platform” | think open sharing |
| “federation” | the source system stays external; access is governed through Databricks |
The outward-facing verbs sound similar, but the control model is not:
If you ignore where the data lives, the answer choices start to look interchangeable when they are not.
| Trap | Better rule |
|---|---|
| treating sharing and federation as synonyms | they solve different boundaries |
| choosing the consumer mode without considering governance | DE-PRO wants secure sharing, not just easy access |
| assuming every external-access question is Delta Sharing | some are federation questions |
| Scenario clue | Stronger answer shape |
|---|---|
| “another Databricks deployment should consume live governed data” | Databricks-to-Databricks sharing |
| “external platform should consume Databricks data” | open Delta Sharing |
| “query external database through Databricks controls” | Lakehouse Federation |
| “source system remains outside Databricks” | federation, not sharing |
This objective usually tests whether data is going out of Databricks or being queried from outside systems into Databricks. If you are publishing Databricks-managed data to external consumers, think Delta Sharing. If you are governing queries against external sources from Databricks, think Lakehouse Federation. DE-PRO often penalizes mixing those directions up.