Our latest Reltio Community Show explored Reltio Data Sharing with Databricks—and how data teams can get trusted, analytics- and AI-ready data into Databricks through delta-sharing, without building and maintaining fragile pipelines or copying data across environments. Led by Ankur Gupta, Principal Product Manager, the session walked through the “why,” “how” and the architecture under the hood, and a practical demo showing just how quickly teams can get started.
Why data sharing (not data movement) matters
Ankur opened with a familiar challenge: fragmented, duplicated, and inconsistent data makes it hard to trust analytics and AI/ML outcomes. Reltio Data Cloud helps unify data across sources and deliver golden, trusted profiles—while Reltio Data Sharing for Databricks makes that trusted data available in Databricks through delta-sharing, without exporting or copying it.
What you can share into Databricks
Reltio Data Sharing makes multiple Reltio objects available to Databricks in near real time, including:
These objects enable data teams to activate BI, data science/ML, and GenAI workloads inside Databricks providing faster TTV and lower TCO.
Key capabilities covered in the session
Zero-copy data sharing
Data is shared—not exported—through delta sharing so teams can access trusted Reltio data in Databricks without building pipelines or managing additional infrastructure.
Fast setup and lower operational burden
Ankur showed that setup is intentionally simple: provide your Databricks sharing identifier, configure the share, and accept/mount it in Databricks. The goal is to help teams go from “need data” to “querying data” in minutes—without weeks of integration work.
Schema options that match how you work
Reltio supports two sharing modes:
-
OV-only sharing: Shares operational values in a simplified, tabular schema (each attribute becomes a column).
-
All-values sharing: Shares OV + non-OV values, keeping a hierarchical structure (in a column struct format).
Both modes can coexist for the same tenant, so different teams can choose what best fits their use case.
Demo highlights: From setup to dashboards (and live updates)
The demo showed end-to-end setup and consumption:
-
Create a new data share in the Reltio console .
-
Copy/paste the Databricks delta sharing identifier.
-
In Databricks, accept the share and mount it to Unity Catalog.
-
Build Databricks dashboard on top of the shared tables and leverage its OOB capabilities.
-
Use Databricks Genie to generate visuals (e.g., pie charts) based on shared data.
To prove “near real time” in practice, Ankur updated a product attribute in Reltio, then refreshed the Databricks dashboard—showing the change reflected in near real time.
Q&A takeaways: scale, control, and governance
A few audience questions surfaced practical considerations:
-
Scale: The approach supports large volumes (e.g., millions of profiles). Initial setup may take longer the first time because existing tenant data must be processed into Delta Lake tables; subsequent updates flow through in minutes.
-
Pause/resume: Temporarily disabling a data share is supported today via APIs (UI controls are not currently available).
-
Schema evolution: Adding new attributes later is supported—new attributes can propagate without manual rework.
-
Security: Data is protected in transit (e.g., TLS 1.2+), access is token-based, and data access is mediated via short-lived signed URLs. Deleting the share revokes access and cleans up associated objects.
Conclusion
This community show reinforced a simple idea: trusted data is the foundation for analytics and AI—but how you deliver it matters. With Reltio Data Sharing for Databricks, teams can publish trusted Reltio profiles into Databricks quickly, reduce operational burden, and accelerate time to insight—without creating yet another integration to maintain.
#CommunityWebinar
#Featured
#DataCloud