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

[joarobert]

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, where you make some model changes in the Direct Query connected dataset (which you may also call the original or main dataset). This may be new measures or dimensions, maybe you are just hiding or unhiding some dimensions, maybe there are some updated power query steps, etc. Any changes to the model in Power BI Desktop.

When you then publish that dataset to the Service, then it would be nice if the changes were automatically added to the composite model also. 

But instead, when we do it, we see that the composite model is not being updated with these changes.

The question, is there any way to update the composite model with the changes in the Direct Query (main) dataset? Or if want to include these changes, do you need to create a composite model from scratch from the updated main dataset? 

Kind regards,
Joachim

[v-xiaoyan-msft]

Hi @joarobert ,

Does the replies above solve your problem? If it has been solved, please mark the correct reply as the standard answer to help the other members find it more quickly.Thanks in advance for your kind cooperation!

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.

[v-xiaoyan-msft · Accepted solution]

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.

[CyprienB]

Hello, 

I have done all these steps but still it is not working. 

To recap I have 2 Dataset A1 and A2 both connected to a sql server with a gateway properly setted. 

I have then a composite Dataset B connected in Direct Query to A1 and A2 ( no need of gateways then)

I do then :

1. Make some changes on desktop on Dataset A1 (ex: adding a measure)

2. Publish the new Dataset  A1 with the same name.

3. Refresh Composite Dataset B directly on Web Service

4. I cannot see the changes on Dataset B

I need to re download B on desktop, update on desktop and then republish. 

Is there a way to avoid that ? 

Thanks a lot in advance, 

Cyprien 

[joarobert]

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...

[GilbertQ]

Hi @joarobert 

What I would suggest is updating to the latest version of PBI Desktop because now when you connect to DirectQuery you can select which tables and or databases you want to include in your selection?

Power BI October 2021 Feature Summary | Microsoft Power BI Blog | Microsoft Power BI