Power BI Premium (PPU) & Synapse (DWH) - Best practices

Question

Hi Team,

We are doing the evaluation of the different ways that we could connect our Synapse (DWH) to Power BI Premium (PPU) and avoid problems with the performance in the future. We want to make sure that we are using the extra features that Power Premium (PPU) offers and the best practices. We are not using AAS.

We would like to know the best practices in terms of type of connection (Import/DirectQuery) and:

• Active Large datasets in Power BI Premium
• Using datasets or dataflows
• Use Normal view or materialised views in the DWH of dimensions
• Use Result-set caching

We know the use of DirectQuery and Normal views is an antipattern according to power-bi-professionals-guide-to-azure-synapse-analytics_2020_06.pdf. We think we should use Materialised Views or Result-set caching. -> What is the recommendation, here?

We were wondering if we use Import as a connection it could be good to use normal views or Materialised Views. We think normal views or a flat table which join the dimensions and fact tables could be fine, because we are going to have the data in the dataset in Power BI -> What is the recommendation, here?

We were wondering when is recommended to use Result-set caching.

We think we are going to datasets. Because we are doing the data preparation in Synapse. Also, dataflows can introduce significant risks to the scalability and sustainability of the solution according to this doc: power-bi-professionals-guide-to-azure-synapse-analytics_2020_06.pdf -> What is the recommendation, here?

Thank you so much!!
Anaid