Databricks Dashboards: Is there an equivalent to Power BI's SEMANTIC MODEL ?

[lucca_luna]

Hello everyone,

I am trying to create a cumulative support ticket chart. I need one line representing the cumulative sum of tickets by creation date, and another line for the cumulative sum by resolution date.

While it's easy to achieve this by creating a new dataset, the dashboard's cross-filtering stops working when I use multiple tables. Is there a way to build a star schema within Databricks dashboards? Alternatively, is there a function to dynamically change the X-axis ?

[-werners-]

metric views in databricks.

https://docs.databricks.com/aws/en/metric-views/

[Ashwin_DSA]

Hi @lucca_luna,

Great question. The cross-filtering issue you're hitting is because the dashboard treats separate datasets as independent. The workaround is to reshape your data so both date dimensions (creation and resolution) live in a single column. You can do this with a SQL view that uses UNION ALL to unpivot the two dates into an event_date column with an event_type discriminator ('Opened' vs 'Resolved'). That gives you one dataset, one X-axis, two lines, and cross-filtering stays intact.

If you also want reusable measures (similar to Power BI's semantic model), take a look at Metric Views in Unity Catalog. They let you define dimensions and measures once, then query them from any dashboard, Genie, or SQL query. 

Your question actually inspired me to write a full walkthrough covering both approaches with a support ticket scenario here. I hope it helps you. 

If this answer resolves your question, could you mark it as “Accept as Solution”? That helps other users quickly find the correct fix.