Forum Discussion

joarobert's avatar
joarobert
Frequent Visitor
4 years ago
Solved

Composite Models: How to apply changes in the DirectQuery-connected dataset?

Hi,  With composite models, you have to change a dataset to Direct Query, which can then be connected to other imported data sources.  This all works fine for us.  However, consider the case...
Need Help
  • Anonymous's avatar
    Anonymous
    4 years ago

    Hi joarobert ,

     

    First make sure that you have checked the features DirectQuery for PBI datasets and AS. in the image below in Power BI desktop.

     

    After published the report,

    Step 1: Go to Settings -> Datasets. Select the dataset that needs to be scheduled.

     

    Step 2: On-Promise sources (both import and direct query mode), you need to configure the On-Premise gateway and add a connection for each data source. Enable the applicable Gateway for your dataset,

     

    Step 3: At first publishing of the report, you need to provide data source credentials with details such as authorization method, privacy level.

     

    Step 4: Go to Scheduled refresh. Enable the Schedule refresh option and create schedules as per your requirement.

     

    Please do make sure that your gateway and data source is up on running during the refresh. With the above steps, you can easily schedule refreshes for your reports that use a composite model. 

     

    For more information,please refer to:Power BI Tips: Schedule Refresh on Composite Models 

     

    Hope it helps,


    Community Support Team _ Caitlyn

    If this post helps then please consider Accept it as the solution to help the other members find it more quickly.

     

     

     

joarobert's avatar
joarobert
Frequent Visitor
4 years ago

Update: I just confirmed that it works when the original dataset is updated with a normal publish to Power BI Service. The Direct Query connected dataset is automatically updated. 

We still have a case, where it does not work. This is a case, where a pipeline updates the original dataset rather than a normal publish as described above. Hopefully we can make it work by making some changes to the pipeline...