Forum Discussion
DAX measures Clean-up in data model?
- 7 months ago
Hello MacJasem,
If you want to clean up unused measures, calculated columns, and tables in a large Power BI model, doing it manually is painful. Fortunately, there are tools that can help:
• Measure Killer – An external tool built specifically for this. It scans your PBIX file and shows which measures, columns, and tables are actually used in visuals. Anything not referenced is flagged as unused, so you can safely remove or hide them.
Official pages:• Brunner BI: https://en.brunner.bi/measurekiller A
• Microsoft Fabric Community blog (correct link): https://community.fabric.microsoft.com/t5/Power-BI-Community-Blog/PowerBi-Measure-Killer-external-tool-to-clean-model/ba-p/3896349 B• DAX Studio: Widely used by developers for deeper analysis. Connect to your model, list/export measures, inspect dependencies, and analyze performance. Official docs: https://daxstudio.org/docs/intro/ (use alongside the DAX reference on Microsoft Learn).
Practical tips before cleanup:• Backup first: Save a copy of the PBIX before changes.
• Hide, then delete: Hide suspected unused items initially to avoid breaking indirect references.
• Iterate and test: Clean in small batches and validate visuals, slicers, drill‑throughs, and relationships after each round.
Direct answer: start with Measure Killer to quickly surface unused objects, then use DAX Studio for validation and performance tuning. The Fabric Community post above walks through setup and options, including MSI/Store/portable installs B. The official Measure Killer page lists features like dependency trees, documentation export, and best‑practice checks A.
Hi MacJasem ,
Thank you for engaging with the Microsoft Fabric Community. In addition to FBergamaschi response, it’s important to note that Power BI does not currently provide a built in way to identify unused DAX measures, calculated columns, or calculated tables at the semantic model level.
For model level cleanup and governance, using external modeling tools such as Tabular Editor is recommended. These tools can analyze dependencies directly within the semantic model and help identify objects with no references. This approach is safer and more scalable, especially when working with large or shared models.
Hope this helps. Please let us know if you need any further assistance.
Regards,
Yugandhar.