Explicit Measures Podcast 454: Looking at the Data Warehouse Roadmap
Aug 29, 2025
Brad Schact, a Microsoft expert on Microsoft Fabric Data Warehouse, joins the hosts for an engaging discussion on the latest updates in data warehousing. They dive into innovative concepts like materialized lake views and their benefits for data management. The conversation highlights user-defined functions in Python, exploring their efficiency and scalability challenges. They also preview an upcoming user group event in Chicago, emphasizing community engagement and the importance of user feedback on future developments. Humor and camaraderie shine through as they tackle complex topics.
AI Snips
Chapters
Transcript
Episode notes
Materialized Lake Views Save Cost And Time
- Use materialized lake views to reduce recompute and manual coding across Lakehouse layers.
- Schedule refresh frequency to balance cost and freshness so you only compute when needed.
Precompute Over Recompute In Cloud
- Cloud storage is cheap so precomputing and storing results often yields better performance and lower cost than recomputing repeatedly.
- Cross-language materialized views (PySpark and SQL) bridge teams and workflows for faster collaboration.
Prototype In Notebooks, Then Promote UDFs
- Test UDFs in notebooks then promote proven logic into persistent user-defined functions for reuse.
- Prefer UDFs for small, repeatable tasks to lower CU cost, but validate scale and governance first.