(00:00:00) The DAX UDF Dilemma
(00:00:32) The Context Transition Trap
(00:00:47) VAL vs XPR: The Core Decision
(00:01:39) The Best Customers Example
(00:02:52) When to Use VAL and XPR
(00:04:54) The Context Transition Problem
(00:05:57) Fixing the Context Transition Trap
(00:08:59) Materializing with Add Columns
(00:13:06) Parameter Types and Casting
(00:16:12) Authoring Checklist for UDFs
In this episode of M365.fm, Mirko Peters shows why most DAX user‑defined functions fail quietly — not because the math is wrong, but because parameter modes, context transition, and materialization are misunderstood.
WHAT YOU WILL LEARN
- The real difference between VAL and EXPR: pass‑by‑value vs pass‑by‑expression and why it changes when your logic runs
- Why using VAL for context‑sensitive metrics freezes results and produces “comfortably wrong” numbers
- How EXPR behaves like a measure, and why you must wrap it with CALCULATE inside iterators to respect the current row
- How to stop recomputing expensive expressions by materializing them once with ADDCOLUMNS and reusing the column
- How data types, coercion, and BLANK handling can quietly change your results in UDFs
- A practical “Mode → Move → Make” checklist you can apply to every new DAX function you write
THE CORE INSIGHT
DAX UDFs are not magic; they are just DAX with sharper edges. VAL vs EXPR decides whether you pass a frozen scalar or a living expression; context transition decides whether row context becomes filter context; materialization decides whether you pay the same expensive cost hundreds of times or once. If you ignore those three decisions, your UDFs will work in demos and betray you in production.
WHO THIS EPISODE IS FOR
This episode is ideal for Power BI developers, DAX authors, semantic model owners, and anyone building reusable calculation logic in enterprise models. If you’ve ever had a “correct” UDF that fails on slicers, ignores the current row, or suddenly becomes slow at scale, this conversation gives you the mental model and patterns to fix it.
ABOUT THE HOST
Mirko Peters is a Microsoft 365 and data platform consultant focused on building trustworthy, reusable semantic models on Power BI and Microsoft Fabric. Through M365.fm, Mirko shares practical DAX patterns, performance lessons, and modeling approaches that help teams move from fragile measures to robust, well‑behaved UDFs that stand up under real workloads.
Become a supporter of this podcast:
https://www.spreaker.com/podcast/m365-fm-modern-work-security-and-productivity-with-microsoft-365--6704921/support.
