Forum Discussion

mp390988's avatar
mp390988
Icon for Post Partisan rankPost Partisan
1 year ago
Solved

Power BI slow running report

Hi,

 

Can someone please explain to me if a report user has a slow running why is it that you need to optimise the data model (i.e. semantic model)? The repor user doesn't interact with the semantic model. They only look at the report.

 

Sorry if this sounds daft but I just can't get my head around this.

 

Thanks. 

  • when you say queries get sent to the underlying database, th vertipak engine, are you talking about Import mode and not DQ?

  • lbendlin's avatar
    lbendlin
    1 year ago

    if everything is DQ then you need to optimize the DQ source for the typical queries from the visuals. For SQL Server that would mean reading the execution plan, updating the statistics and tuning the indexes.

7 Replies

  • Every time the user interacts with the report, new queries for all affected visuals are run against the semantic model. You can visualize that by opening the "Performance Analyzer"  window and starting the recording  on the Power BI Desktop. Pretty much the same thing happens on the service (against the semantic model in the workspace).

  • Deku's avatar
    Deku
    Icon for Super User rankSuper User

    When a viewer opens a report each visual needs to send queries to the underlying database, the vertipak engine, to order to get the data used in the visual. The time taken to answer the query will depend on the size of the database, and the complexity of the query, which is influenced by the design of the data model and the measure definitions

    • mp390988's avatar
      mp390988
      Icon for Post Partisan rankPost Partisan

      when you say queries get sent to the underlying database, th vertipak engine, are you talking about Import mode and not DQ?

      • Deku's avatar
        Deku
        Icon for Super User rankSuper User

        In import the storage engine is veripak, for direct query it would be some remote database