Forum Discussion
Gateway with different creds to Datasource
Easiest way is to think of it like this:
While you are in Power BI Desktop, you have some locally stored credentials that are used to make the connection. If you save your PBIX file and pass it to someone else, the credentials are not stored in the PBIX file itself, but in the local cache of your Power BI Desktop application. This new person would need to enter their own credentials to access the DirectQuery data, or run a refresh (Import).
When you publish to the Power BI Service, again no credentials are sent along with the file. Since the cloud is outside your network, unless your data source is also in the cloud, it has no direct connection to your data source.
An On-Premise Data Gateway acts as a secure bridge. It gets installed on hardware inside your network and you make a specific connection to the Office 365 tenant. Under "Manage Gateways" in the Power BI Service, you can add Data Sources, their credintials, and which users can access this data source. This is now avaiable to be used by any project that contains that data source and is being configured by someone on the access list. (small note, a project that wants to use a gateway needs to ensure that every data source is within the gateway, or is a cloud data source).
When you go into the "Schedule Refresh" of your Power BI Dataset on the Service, you must select a gateway. If its configured correctly you will see it appear in the gateway listing. Whenever this refreshes (import) or connects (direct query), it will talk only to the gateway. The gateway itself will handle the authentication and passing the data between the cloud and your network. The report itself doesn't need any credentials.
The credentials while in the cloud are all managed via "Manage Gateways" and work irrespective of how you publish a project/dataset into the Power BI Service.
When configuring the gateway itself, you'll only need to specify the Server Name, the database name and the appropriate log on credentials.
As for SQL. You can specify the SQL statement as part of your importing, there is an "Advanced options" down below. You won't need to place the SQL into the gateway config.
Sorry maybe I was not clear.
If I add a SQL statement through the spot you just showed me then I cannot add different logon Creds even though it says that it’s required.
If however I just skip the SQL Query then all is good, I get prompted to change the creds and then can see all tables.
Is there a way to just save the Server and DB with the correct creds, and then run my queries on top? I have been adding additional queries for my different sections (i.e. for my Service Desk tool, I have one for Tickets and one for CMDB) as this was the only way I could work out how to join tables (only way to solve this Problem)
Power BI Desktop remembers credentials you have used, which might be why you aren't getting the credentials message. You can change the credentials by clicking on 'Edit Queries drop arrow' -> 'Data Source Settings'
Credentials are not saved into the project itself. Whatever you choose in the Power BI desktop has no bearing on your gateway.
If you do want to modify what is stored in Power BI Desktop you can try:
If you are using a file that is already connected to the data source, you can select it and choose "Edit Permissions".
If you are in an empty project, you can select at the top "Global Permissions" and find your data source. This will let you update the permissions.
This is what i get when trying to use domian admin creds when connection to a data source.
Power BI running as domain user (run as admin) and datasource has my domain admin creds.
If you select 'Edit Queries drop arrow' -> 'Data Source Settings' and then edit the permissions. Is your data file set to this:
That is what Power BI desktop is complaining about.
Alternative you could instead use an SQL Authentication (database username and password) and avoid it entirely.
Based on your description it sounds like there is a specific account that you need to use, so if the user account that is running Power BI Desktop is different to the one in the database, you could try holding shift before right-clicking on your Power BI desktop icon and selecting "Run as different user".