Data views and virtual tables
- Last UpdatedOct 22, 2025
- 1 minute read
Data views enable you to access subsets of data items from CONNECT data services in data-driven applications, where the items can be used for data science enablement. With data views, you can bridge your raw CONNECT data services data to third-party applications like Microsoft Power BI, where it can be used for analytics, machine learning, and so on.
Data views
Performance metrics
|
Performance metric |
Value |
Context |
|
Max streams per data view |
100,000 |
Hard-coded limit. |
|
Response times |
Variable |
Dependent on stream count and data density. |
Additional notes
-
The number of data views per tenant or namespace is generally not a limiting factor.
Virtual tables
Test bed description
-
Testing included long-running jobs and streams configurations.
Performance metrics
|
Performance metric |
Value |
Context |
|
Stream support |
No hard limit |
Dependent on configuration and data rates. |
Additional notes
-
Performance varies based on stream configuration, data density, and the region of the Databricks deployment.
-
Configuration guidance:
-
Stored vs interpolated data impacts volume and performance.
-
Narrow vs standard mode affects table structure.
-
Best performance usually achieved below 100,000 stream references.
-
Initial virtual table creation time depends on history length.
-
Re-creation due to changes to schema or data may result in long-running jobs.
-