Hello Amit! The only cost estimator I know is this one ---> Pricing Calculator | Microsoft Azure For any other information regarding a transition to Microsoft Fabric, there are so many variables to take into account, and it is hard to define common guidelines 🙂 Let's wait for other comments and see if someone has some helpful info!

I think the new DP-600 exam was meant to replace the old (and unavailable) DP-500 exam (Azure Enterprise Data Analyst Associate Certification). As far as I remember, the DP-500 was also centered around Power BI and some advanced stuff like Tabular Editor, XMLA Endpoint, semantic modeling, etc. along with questions regarding how to design and implement enterprise-scale analytics solutions using Microsoft Azure (in particular, Microsoft Purview and Azure Synapse Analytics) and Power BI (Desktop and Service) So I think that's the reason why Microsoft has included a lot of questions about Power BI in the DP-600 exam 🙂

Hi Jerry, like you, I consider semantic models as data marts in some ways because I can create different semantic models for specific areas using tables residing in the same database. In the same way, each model can be customized based on the analytics needs of each area. In my opinion, Data marts in Power BI Service can be considered as a preliminary small version of Fabric Data Warehouses 😄

@Jerry Lee it is exactly what me and my team suggested (and did) for a project before Fabric: developers for each department/area created their own semantic models to support their analysis 🙂

Hello Surm! In my opinion, having a general understanding of both Power Query and Python can be the best option but, I understand your point 🙂 Based on my experience, Python is far more powerful than Power Query when it comes to heavy and complex data manipulations and transformations. In this context, Fabric notebooks have the Data-Wrangler which can help non-expert people to apply transformations on data without writing any Python code. Of course, the data wrangler is NOT like Power Query and right now it can be used only for data preparation (i.e. cleaning, manipulation, ...). I hope Microsoft will improve this to have a complete Python experience to handle complex ETL processes because, in my opinion (and based on my personal experience), Power Query is ok but, it is not suitable for complex operations (to me, it cannot be considered a 100% real ETL tool)🙂

Hello Tobias, based on the Microsoft documentation and suggestions, and based on my personal experience: - Data ingestion: the tool to use depends on what you want to achieve. Pipelines are great if you want a simple and quick copy-data activity (source --> Fabric). However, if data transformations are involved, Dataflows and Notebooks are the best choices. Mind that, based on my experience, Spark Notebooks are way more powerful than Power Query when it comes to complex transformations, table joins, etc. - Medallion Architecture: the answer is "it depends" 😄 it depends on the data journey you want to set and how many layers you want to build to prepare data for analysis and reporting. In other words, the "classic" three-medallion architecture (as also suggested by Microsoft) can be further customized depending on several factors (e.g., what are the data validation rules? Do I need to create different Lakehouses for each department? Do I have different rules for each department data?). Also for what concerns the tools to use between each layer, I would say the answer is "it depends" (e.g., they can be notebooks, stored-procedures - if you go with Warehouses). In general, I would say that there are no straight answers when it comes to implementing a Fabric solution but you can narrow the circle of options by collecting information on what you need to do on data 🙂 What do you think?