Performances of SQL Server VS Analys Services in import mode

[MaxItaly]

Hello everyone,

my question is quite simple:

I have to analyze a database with Power BI Desktop in import mode: does the fact that the database has been built and hosted in Analysis Services lead to better performance than the same database built and hosted in SQL Server (Cubes VS Tables)?

I don't know how Power BI works in detail, but I have this doubt that Power BI downloads the data and re-optimizes it in its own way, nullifying the utility of the host of the data.

Does anyone have a more in deep knowledge about this?

[Anonymous · Accepted solution]

If you are about to consider Power BI Desktop Import mode then doesn't matter wheather the Source Database biuld on Analysis Services or SQL Database.

Power BI Desktop imports all the data into the pbix file in Import mode and Optimize the data in its own way and keeps it in memory.  

Data in the cubes are in memory where as the tables of SQL Server are not. 

I think you won't be having Performance issues once the data was imported to Power BI Desktop. Only conern is Processing time (time to import data into pbix file) may vary between SQL Server and Analysis Services because Analysis Services keeps all the data in memory(readily available and fast retrivel) and SQL Server doesn't.

Anyways if are tryng to import huge amount of data into Power BI Dekstop then you may ran into Performance issues, In that case I would recommand to use Analysis Services and then live connceted Power BI with the cube.

[MaxItaly]

Many thanks, you've been really clear. I'm happy to know that I understood correctly how it worked.