DP-600 Fabric Analytics Engineer - how much DAX do you actually need?
I've been using Microsoft Fabric for about 8 months and I'm preparing for the DP-600 exam. My background is mostly Power BI report development and some ADF pipeline work, so I feel okay on the visualization and orchestration side. What I'm trying to gauge is how deep the DAX requirements actually go - I can write solid calculated columns and measures but I've never needed advanced time intelligence or complex CALCULATE modifiers in production.
The exam skills outline lists Lakehouses, Warehouses, Spark notebooks, data pipelines, semantic models, and real-time analytics as core areas. I'm weakest on Spark and real-time analytics - never touched Eventstream in production and my PySpark is limited to basic transformations. I scored 58% on a practice assessment and need 700 to pass.
I'm planning 8 weeks of prep at roughly 1.5 hours daily on weekdays. I've done SC-900 and PL-300 before so I know the Microsoft exam format. Are there specific Fabric features that keep showing up? I've heard OneLake architecture comes up constantly.
The Spark section was harder than the study materials suggest. Spend real time on Delta Lake operations in Fabric notebooks - MERGE statements, time travel queries, and optimization commands. That came up more than I expected from a Power BI background.
DAX on my exam wasn't as deep as I feared. You need CALCULATE, context transition, and basic time intelligence like TOTALYTD and SAMEPERIODLASTYEAR, but I didn't get anything requiring advanced iterators or complex KEEPFILTERS scenarios. Your current level sounds adequate.
Eight weeks at 1.5 hours daily is a solid plan. I went from 61% on my diagnostic to 78% on the actual exam in about 7 weeks. The Microsoft Learn Fabric paths are the best free resource - more reliable than third-party dumps that don't reflect current Fabric features.
OneLake and the shortcut architecture definitely show up a lot - probably 15-20% of my exam was about how data flows through OneLake and why you'd use Lakehouse vs Warehouse for different scenarios. Get that conceptual framework solid before anything else.
Honestly, the DAX requirements aren't as deep as I feared going in. I've got a similar background to yours and I was worried I'd need to be writing complex time intelligence functions from scratch, but it's more about understanding how DAX works within the Fabric ecosystem than memorizing syntax. Things like knowing when to use calculated columns vs measures, understanding evaluation context, and being comfortable with basic filter functions is about the level you need. I studied maybe 45 minutes a day during my lunch break and a couple hours on weekends for about six weeks, and the dp 600 microsoft fabric architecture 4 practice material was actually really useful for calibrating how deep to go on each topic.
Where I'd spend more time if I were you is on lakehouses, the medallion architecture, and how Direct Lake mode actually works under the hood. That stuff came up more than I expected and it's easy to underestimate if you're coming from a pure Power BI background. Your ADF experience will help more than you think on the orchestration questions too.
I failed it in March and passed in May, so maybe my experience helps. First time around I did exactly what you're describing, leaned on my Power BI background and assumed the DAX questions would be report-level stuff. They weren't. I got hit with calculation groups, semantic model optimization, and a couple of questions where you had to pick the right DAX approach for DirectLake vs import scenarios. It's not that the DAX is crazy deep, it's that they test it in the context of Fabric architecture decisions, which I hadn't prepped at all.
Second attempt I flipped my prep. Spent way less time on visuals and way more on how semantic models, lakehouses, and warehouses actually fit together. Practice tests helped me find the gaps fast, I used dp 600 microsoft fabric architecture 4 and a few others to drill the architecture side specifically. Honestly with 8 months of Fabric under your belt you're probably fine on DAX syntax. Just make sure you can answer "which storage mode and why" type questions cold, because that's where they sneak the DAX in.