Has anyone had any experience trying to create a 'Golden Dataset' with a super complex relational database? This is more of a general discussion as I was hoping to understand whether people are coming into similar problems and what they decided to do long term. I have done loads of research and I can 100% see the benefits, but it is ALOT of work creating the Golden Dataset given our data model and then even more re-work with our BI reports. We have so many tables to combine with millions of rows per table. Combining them all into a single table (or a few tables where possible) does not seem feasible. I am wondering whether people have had the same problems and how they have overcome it. Is worth it in the long run? If someone has already done this with a complex relational database, it would be great to hear your learnings from the experience.

SoniaB102 doesn't matter how many tables or large your database is, as a best practice you want to thrive for Star schema in Power BI regardless of how the table is maintained in the backend. I have done many large projects with a lot of rows (size of database) and the golden dataset worked perfectly, never ran into any issue until now. Now the composite model is available, for some ad-hoc requirements, maybe it will be easier to link other data sources with the golden dataset. I will highly recommend using this approach. End of the day, it is easy to manage, with better security, few people are allowed to touch the model, make changes, add measures, etc. I find it a win-win for my projects. This is my experience. I'm sure others have different thoughts/experiences.