Modeling and Shared Datasets

[PBCIT]

Ok here is a question for you all, and I'm sure there is a good solution that I am just not seeing. Currently, I've started utilizing Power BI Shared Datasets and I love it, they've been a lot easier to manage. I use an on-prem data gateway on a virtual machine and I have all my different sources in it. Here's the question though. There are tables in a database I use that I've found I use in many other datasets, but there are also additional sources that are only used for a particular report/app. I don't like seeing my data duplicated in other datasets, however I don't know if I should consolidate and add all the sources to the one shared dataset as I don't want that shared dataset to become too large (not to mention the relationships between the different sources can get a little dicey and I usually have to be creative to get them to model together and I worry putting those together would ruin my data model schema). Is it better to have all of the data sources in one large dataset, or is it better to try to condense it down as much as possible and if there is a unique source for a report, create a separate dataset for the report to save on dataset capacity and to opt for simpler schemas?

[amitchandak · Accepted solution]

@PBCIT , I think division should be based on work Area. If you are on premium, you might like to have one large dataset. Other keep it work Area wise.