Hi everyone!
I’ve just released an open-source tool that generates a semantic layer in Databricks notebooks from a Power BI dataset using the Power BI REST API. Im not an expert yet, but it gets job done and instead of using AtScale/dbt/or the PBI Semantic layer, I make it happen in a notebook that gets created as the semantic layer, and could be used to materialize in a view.
It extracts:
- Tables
- Relationships
- DAX Measures
And generates a Databricks notebook with:
- SQL views (base + enriched with joins)
- Auto-translated DAX measures to SQL or PySpark (e.g. CALCULATE, DIVIDE, DISTINCTCOUNT)
- Optional materialization as Delta Tables
- Documentation and editable blocks for custom business rules
🔗 GitHub: https://github.com/mexmarv/powerbi-databricks-semantic-gen
Example use case:
If you maintain business logic in Power BI but need to operationalize it in the lakehouse — this gives you a way to translate and scale that logic to PySpark-based data products.
It’s ideal for bridging the gap between BI tools and engineering workflows.
I’d love your feedback or ideas for collaboration!
..: Please, again this is helping the community, so feel free to contribute and modify to make it better, if it helps anyone out there ... you can always honor me a "mexican wine bottle" if this helps in anyway :..
PS: Some spanish in there, perdón... and a little help from "el chato: ChatGPT".
- Marvin
Marvs
