Multiple PBIX files or a tabular analysis cube?

[haassiej]

Hi everyone,

I have a question regarding the use of separate PBIX files or the use of a Microsoft Analysis Service CUBE.

I work with various financial subject areas. As an example: Cost Analysis, Holds etc etc.

And now I am transferring this to Power BI. Now we have two choices that we can make regarding the chosen technique:
 
I. For all subject areas, I create a separate PBIX, connected to the database and then the choice for import or direct query. Then the users start with these separate (published) PBIX files and can optionally choose to analyze in Excel via the web version. This would amount to a lot of (+ - 30 pieces) loose PBIX files

II. We start with (Azure) Analysis Services and make a Tabular model per topic area, with which we connect with Power BI and / or Excel.

Now I wonder what you recommend from my own experience? 🙂

And what are the pros and cons?

[v-yuezhe-msft · Accepted solution]

@haassiej,

Do you use Azure Analysis Service? If that is the case, please note that refresh model on Power BI Service with Azure Analysis Service data source in import mode is not supported. If you want to refresh the dataset in Power BI Service, you would need to use "Connect Live" mode to connect to Azure AS in Power BI Desktop.

Moreover, personally, I prefer to have a single Tabular model and create report connecting to the model in a single PBIX file, which is more easy to manage comparing to multiple PBIX files.

Regards,
Lydia

[haassiej]

@v-yuezhe-msft For refreshing the dataset i could use the gateway, right? If I decide to use the import mode.

[v-yuezhe-msft]

@haassiej，

For Azure Analysis Services data source, gateway is not required when you refresh the dataset.

Regards,
Lydia