if you are using power bi to develop dashboards from a cube, do you think its better to create the cube in tabular rather than MDX and why?
currently we are creating a cube in via direct querry mdx, and ive already noticed that i can't create measures in in power bi but i am able to do that in tabular.
Well that question totally depends on what features you need to build the cube. There is still some very nice features in Multidimensional that isn't available in Tabular. But if you can build the cube you need in Tabular then I would go that way. Power BI's own models are build on the same engine as Tabular models hence there is a better "understanding" between Power BI and Tabular than Power BI and Multidimensional.
A few weeks ago I attended a session with Chris Webb where he talked about this question and he listed several things that isn't supported in Power BI when using direct query to Multidimensional, but also stated (and I agree with this) if you already have a Multidimensional cube you should only switch to Tabular if there is things you need to do in Power BI that is not supported for Multidimensional. However if you want to start over and build a new cube I would go with Tabular unless there is features that you need that is only supported by Multidimensional.
@sdjensen ps do you know where i can find what is not possible to do in MDX direct query vs Tabular direct query?
Here is some of the things that currenly doesn't work with live connection to SSAS MD:
SSAS MD features that isn't currently supported by Power BI:
@sdjensen oh wow thats a big one, role based security doesn't work with mdx and direct query in power bi?
My post didn't mention RLS, but you can read about the limitations about RLS here: https://docs.microsoft.com/en-us/power-bi/service-admin-rls#limitations
RLS has never been used for live connections to Analysis Services this also apply to Tabular.
"Only ETL, and DirectQuery connections are supported. Live connections to Analysis Services are handled in the on-premises model."
@sdjensen yes the scenario here is to implement role based security in an MDX cube, using a direct query connection on Power Bi Report Server, sorry i dont quite understand what you saying it does work or it doesn't? will check out the documentation
@sdjensen also what is ETL?
extract transform load (ie imported mode?)
An ETL source could be a SQL database where you import data into your model. You can't user Power BI RLS for live connection to cubes - here security is handled with roles in the cube.