I have watched this video twice. This is a must watch for anyone dealing with large datasets. In my case everything else was fine except currency col. Which alone was consuming 63% of the total. So, converting it Fixed decimal reduced it down to 43% and Model size by over 20MB. Thank you! But I do have a couple of questions: 1. Besides reducing Model size what other benefits are there? 2. Vertipaq Analyzer summary shows Model size 119 MB but the size of PBIX appears to be 37 MB whereas on Power Bi Service the same dataset appears to be 47 MB. Any idea what is this difference about?

Let's say we need to keep the ChatID column to be able to identify unique chats (no foreign key). Would an alternative be to use a dense ranking function to create a surrogate key to use a distinct measure on in PowerBI? Might be another tip? Had an issue recently with users vs sessions (unique sessions vs sessions). EG: ChatID -> SK Key A123 -> 1 A123 -> 1 B122 -> 2

video was long but it was worth watching it! Question though: I have a large dataset size: 0.86 MB. In my opinion I have taken care of the mentioned points here, but I will inspect more. However, I am afraid this dataset will get higher than 1 gb size in a month's time. I currently have a Pro workspace but can upgrade to premium per user. Do you think upgrading to premium per user can resolve the 1 gb limit? If I am not wrong Premium per user allows a model size of 100 gb. Also I am a bit concerned about user. Will all existing users be able to still view the report if I upgrade to premium per user? Thanks