How do you handle RLS when dealing with a composite model? Direct query for detailed fact data (RLS at source) + import mode for aggregated fact data + dual storage mode for dimensions.
We tried that in our project earlier, defined RLS on aggregated data and used SSO for detailed direct query but SSO didn't work with that. So had to create 2 models, direct query and import separate.
You need a gateway to setup since this is direct query. When I setup a gateway, I’m not getting an option to edit credentials since it says admin has granted access credentials are not required. What to do ? Please help.
Hi ! Thanks for your video. Some weeks ago we created a report using Power BI Report Builder that connects to a data source which is an Analysis Services model in Azure. That model have roles that secure the data for some users. But, when we publish the report in the Power BI service, the report refreshes using the account configured in the connection, but there is no option for Power BI to use the user that is using the report to connect to the model instead the one that is configured in the connection, so, as a result, the RLS doesn't apply in this case. Any clue to solve this issue?
Hi Patrick. I have a quick on RLS. I have set up RLS within my PBI desktop and everything seems to work when check the "View As" or "Manage Roles" and I have this dataset published to PBI Service. I intend to have people connect to the published dataset from EXCEL but they are only to see the data that is designed or set up in the RLS. Is that possible? Do you have a vid on such that I can wach?
Hi' I would love to know more about composite model and RLS. I have a composite model with RLS, some of the Facts are in direct query and some are not. The RLS works well when it comes to slicing the data but creates unusual slowness in real time reports. I would be happy to help on this matter
I have a dataset where incremental refresh was setup and running for quite long time.All of a sudden i have included a relationship with a new calculated column (created using tabular editor). Is there any way to refresh the dataset even with partition(process full)??? Im getting failure even with one day partition(process full), as it needs to recalculate the new column and relation..Can this issue be fixed any chance???
Hi patrick, Hi Adam, how am I able to publish a 1.2GB report file into my Power Bi service with PRO licencing when the maximum dataset limit is 1GB am I missing something?
Hi Anime, Please add date parameter to your report and try to get the data less than 1 GB. Publish it into power BI service and change the date parameter again.
@@SumanSharma-fd5mg Hi Suman, thanks but my question is why I'm not getting an error message that I have exceeded 1GB data size limit? If the pbix file is above 1GB it should not be able to publish in service with pro licensing right? Since the maximum dataset limit for PRO is 1GB So why am I not getting that error? Am I missing something? I was even able to publish report with 1.3GB into service
I am trying sort of the other way around, I have one source + dataset with NO RLS set up, but I want to build one report based on that WITH RLS, I have been trying to create a composite model with direct query to the original dataset and import only the dimensional table that I want to use for rls. I think that the "weak" relationship between the imported dimensional table and the dq fact table is somehow not working, the filter does not propagate, perhaps worth investigating for another video? :)
Hi, I have a requirement where I need to manage a lot of users in different roles in RLS. At this point I manually update them in Security -> Manage Roles. Is there a way instead of manually maintaining the users for the roles in RLS? One other way is creating Distribution lists but in my requirement, creating a Distribution list is not an option. Please let me know in case there is any cool solution for this.
In the Menu where you create the Filter for the roles. You can use dax Formular Username() . So you can Work with an Filter Expression Like mytable[columnwithuserid] = Username(). There ist another function Like userprincipalname(),which could Work AS Well.
Dynamic roles can be created only if we have hierarchy right? My usecase doesn't have hierarchy. The users for Roles are Random and Manually provided in a CSV file.
I dont See a Need forba hierachy I guess the Data in the CSV File is Like User;Group User1;GroupA User2,;GroupB You Just Need to ready in that Data and Connect IT with the Rest of the model
@@bhuvaneshkj9795 make a column in a FACT table with user email addresses, make a measure if contains ( column) USERPRINCIPAL() than 1. Put this measure as a filter and filter it as = 1. thus users will see only rows with their emails.
![[Live สด] การออกรางวัลสลากกินแบ่งรัฐบาล งวดวันที่ 16 ธันวาคม 2567](http://i.ytimg.com/vi/dNTiBBthKgQ/mqdefault.jpg)
How do you handle RLS when dealing with a composite model? Direct query for detailed fact data (RLS at source) + import mode for aggregated fact data + dual storage mode for dimensions.
I'd probably do the RLS as a role on the front end vs. at the source if the model is complex
We tried that in our project earlier, defined RLS on aggregated data and used SSO for detailed direct query but SSO didn't work with that. So had to create 2 models, direct query and import separate.
Thank you very much for sharing it. When I go to settings, the "Data source credentials" is grayed out. Does anyone know what could it be?
You need a gateway to setup since this is direct query. When I setup a gateway, I’m not getting an option to edit credentials since it says admin has granted access credentials are not required. What to do ? Please help.
Hi ! Thanks for your video. Some weeks ago we created a report using Power BI Report Builder that connects to a data source which is an Analysis Services model in Azure. That model have roles that secure the data for some users. But, when we publish the report in the Power BI service, the report refreshes using the account configured in the connection, but there is no option for Power BI to use the user that is using the report to connect to the model instead the one that is configured in the connection, so, as a result, the RLS doesn't apply in this case. Any clue to solve this issue?
i find this so frustrating!
lol nice. no BS moment. just straight to the point. exactly came for this knowledge!
The perfect video, many thanks
For this to work. Do we need to then create a user for each and every person in the underlying data source as well?
Hi Patrick. I have a quick on RLS. I have set up RLS within my PBI desktop and everything seems to work when check the "View As" or "Manage Roles" and I have this dataset published to PBI Service. I intend to have people connect to the published dataset from EXCEL but they are only to see the data that is designed or set up in the RLS. Is that possible? Do you have a vid on such that I can wach?
Team with respect to matrix visual is there a possibility to turn of drill down and keep only focus mode and restrict screenshot for certain users.
Hi'
I would love to know more about composite model and RLS.
I have a composite model with RLS, some of the Facts are in direct query and some are not. The RLS works well when it comes to slicing the data but creates unusual slowness in real time reports.
I would be happy to help on this matter
I have a dataset where incremental refresh was setup and running for quite long time.All of a sudden i have included a relationship with a new calculated column (created using tabular editor). Is there any way to refresh the dataset even with partition(process full)??? Im getting failure even with one day partition(process full), as it needs to recalculate the new column and relation..Can this issue be fixed any chance???
Thanks Patrick!
Hi patrick,
Hi Adam,
how am I able to publish a 1.2GB report file into my Power Bi service with PRO licencing when the maximum dataset limit is 1GB
am I missing something?
Hi Anime, Please add date parameter to your report and try to get the data less than 1 GB. Publish it into power BI service and change the date parameter again.
@@SumanSharma-fd5mg Hi Suman, thanks but my question is why I'm not getting an error message that I have exceeded 1GB data size limit?
If the pbix file is above 1GB it should not be able to publish in service with pro licensing right?
Since the maximum dataset limit for PRO is 1GB
So why am I not getting that error?
Am I missing something?
I was even able to publish report with 1.3GB into service
Hi! Is it possible to configure dynamic RLS on Direct Query Dataset?
Idk how to get RLS done for DQ at source end if the source is some other DB like postgre or Maria
Is there a way to "Test as Role" in the service for a thin report and not just the dataset its based on?
Do any of you know, where I can buy a t-shirt like Patrick is wearing ?
I am trying sort of the other way around, I have one source + dataset with NO RLS set up, but I want to build one report based on that WITH RLS, I have been trying to create a composite model with direct query to the original dataset and import only the dimensional table that I want to use for rls.
I think that the "weak" relationship between the imported dimensional table and the dq fact table is somehow not working, the filter does not propagate, perhaps worth investigating for another video? :)
Great! How does it work with Embedded (master user)?
Hi, I have a requirement where I need to manage a lot of users in different roles in RLS. At this point I manually update them in Security -> Manage Roles. Is there a way instead of manually maintaining the users for the roles in RLS? One other way is creating Distribution lists but in my requirement, creating a Distribution list is not an option. Please let me know in case there is any cool solution for this.
This is probably to unclear to get an proper answer.
Dobyou know that you can create dynamic roles in Power BI with DAX Code?
In the Menu where you create the Filter for the roles.
You can use dax Formular Username() .
So you can Work with an Filter Expression Like mytable[columnwithuserid] = Username().
There ist another function Like userprincipalname(),which could Work AS Well.
Dynamic roles can be created only if we have hierarchy right? My usecase doesn't have hierarchy. The users for Roles are Random and Manually provided in a CSV file.
I dont See a Need forba hierachy
I guess the Data in the CSV File is Like
User;Group
User1;GroupA
User2,;GroupB
You Just Need to ready in that Data and Connect IT with the Rest of the model
@@bhuvaneshkj9795 make a column in a FACT table with user email addresses, make a measure if contains ( column) USERPRINCIPAL() than 1. Put this measure as a filter and filter it as = 1.
thus users will see only rows with their emails.
hi, is office is dead after chatgpt?