Semantic model uses default data connection - semantic model connecting to data warehouse

[torishriver]

I have a semantic model in Power BI that I want to add a table to from a data warehouse. It refreshes on my desktop just fine, but when I publish it to the online version it gives me this error:

We cannot refresh this semantic model because this semantic model uses a default data connection without explicit connection credentials. Please replace the default data connection in the semantic model settings with an explicit cloud or gateway data connection

I've tried using both Windows and Microsoft credentials, as well as each of the privacy levels associated with each and I cannot get the refresh to stick. Please help!

[v-hashadapu]

Hi @torishriver , Thank you for reaching out to the Microsoft Community Forum.

We find the answer shared by @Jai-Rathinavel  is appropriate. Can you please confirm if the solution worked for you. It will help others with similar issues find the answer easily.
Thank you @Jai-Rathinavel  for your valuable response.

[audreygerred]

Hi! I am assuming that your new data source is either on prem or is online but your company is choosing to require a gateway. If you connected to an Excel file using the Excel connector or a csv file using csv/text connector or folder using folder connector these would all require a gateway. You can set up a personal gateway and then your computer will have to be on and the gateway will have to be on for teh refresh in service to work. If you did use one of teh above mentioned connectors, you could swap to the online versions like Web, SharePoint folder, etc.

[Jai-Rathinavel · Accepted solution]

@torishriver To fix this error just follow the below steps

1. Go to Dataset -> Settings 

2. Under the Settings you can find Gateway and Cloud Connections

3. Click Create Connection 

4. Under here you have to create a new connection and provide your SQL Server string and your credentials, authenticate it and click Create (located at the bottom)

5. Once created go back to your dataset settings and under Gateway and Cloud Connections choose"New_Connect" (in my case) 

6. Now you should be able to refresh the model

Thanks,
Jai

[torishriver]

That seems to have worked but I'm now receiving this error-- do you have any idea what to do to fix this? I can't tell if it's having a problem with none vs organizational privacy levels but it's failing to refresh regardless of which privacy level I set the gateway to:

Information is needed in order to combine data. Please specify a privacy level for each data source.;Information about a data source is required.. The exception was raised by the IDbCommand interface.

[v-hashadapu]

Hi @torishriver , Thank you for reaching out to the Microsoft Fabric Community Forum.

When you publish a Power BI semantic model that combines data from more than one source, the Power BI Service enforces stricter rules than Desktop. Desktop lets you bypass privacy settings with Ignore privacy levels, but the Service does not. That’s why you’re now seeing the error asking for privacy levels and credentials.

The first step is to make sure that every data source in your model is properly mapped in the Service. Go to your workspace, open Dataset Settings -> Gateway and cloud connections and confirm that each source is linked to the correct gateway or cloud connection. The connection details, especially server and database names, must match exactly with what you used in Desktop, even small differences like using an IP address versus a server name can break the mapping.

Next, set explicit credentials and privacy levels in the Service. Under Dataset Settings -> Data source credentials, edit each data source, re-enter the authentication method (Windows, Microsoft or SQL login) and assign a privacy level. In most organizational setups, Organizational is the right choice, but you can use Private or Public if appropriate. You’ll also want to edit each data source inside Manage gateways and assign the same privacy level there. Power BI requires that every source has a defined privacy level before it can combine them.

If your model merges or appends tables across different systems, you may still encounter issues. Push those joins into the database itself, for example, create a SQL view in your warehouse that combines the tables and then connect Power BI to that single view. This reduces cross-source privacy checks and usually makes refreshes more reliable.

Finally, if the refresh still fails, check the Refresh history in the Service. The detailed error log will usually point to the specific query or source that’s missing a privacy level or credential. That way you don’t have to guess which one is causing the issue.

Data refresh in Power BI - Power BI | Microsoft Learn

Privacy levels in Power Query - Power Query | Microsoft Learn

Set privacy levels (Power Query) - Microsoft Support

Add or remove a gateway data source - Power BI | Microsoft Learn

View and manage on-premises data gateways - Power Platform | Microsoft Learn