Update July 2020: While this method still works with Power BI Desktop June 2017 release and SSAS 2017, it potentially doesn’t work with the later versions of Power BI Desktop (Jul 2020) and SSAS 2019 as there are new DAX functions that are available in the Power BI Desktop but not in SSAS 2019. So use this method as an experimental method AT YOUR OWN RISK!
Have you created a robust model in Power BI Desktop and you are looking for a way to import it to an instance of SQL Server Analysis Services Tabular? Hmm, it would be highly beneficial if you could import Power BI model to SSAS Tabular and it potentially saves lots of development time and costs. The good news is that with SQL Server 2016 and SQL Server Data Tools for Visual Studio 2015 it is possible. In this post I show you how to import Power BI Desktop model to SSAS Tabular 2016. Unfortunately, you cannot do the job in any prior versions of SQL Server, SQL Server Management Studio or SSDT.
Power BI Desktop is a fantastic report authoring tool. I have lots of experience working with Tableau as well and I can say, man, Power BI is growing very quickly. Lots of awesome ideas have been added to Power BI and a lot more is coming. But, It might be a question for some of you that is that possible to connect to a Power BI Desktop model from Excel, SQL Server Management Studio (SSMS) or SQL Server Profiler? The answer is yes, you can. But, how on earth someone should connect to a Power BI Desktop model from Excel, SSMS or SQL Server Profiler? Well, it could be useful for the following scenarios:
Connecting to the model using SQL Server Profiler for performance tuning, monitoring and so forth
Again, if you have some performance issues you might need to connect to the model from SSMS
You have a complex model and it’s hard for you understand it, but, you are a great Excel developer, so you can connect to Power BI Desktop model from Excel so you can use reach features available in Excel like named sets
Just for curiosity! You are curious about writing MDX codes over an existing model, you want to see how your model look like in Excel and so forth
In this article I show you how to connect to Power BI Desktop model regardless of any use case scenarios. So for whatever reason you’d like to connect to a Power BI Desktop model this post will help you achieve your goal.
How it works
Power BI Desktop uses xVelocity technology and loads data into memory. It uses a local instance of SQL Server Analysis Services (SSAS). It does the job by running msmdsrv.exe file which can be found in “bin” folder under your Power BI Desktop installation folder which is normally under you Program Files. The msmdsrv.exe is indeed the SSAS service file. So even if you haven’t installed SSAS on your machine Power BI Desktop runs msmdsrv.exe. When Power BI Desktop runs msmdsrv.exe it creates a local instance of SSAS. This local SSAS instance uses a random port number so it would be valid until Power BI Desktop is not closed or the msmdsrv.exe is not killed from Task Manager.
So, we have a local instance of SSAS using a random port number. Therefore, we should be able to connect to the instance from Excel, SSMS or SQL Server Profiler only if we know the port number.
Note: If you have installed an instance of SSAS on your machine you can find msmdsrv.exe under “\OLAP\bin” folder from SQL Server installation path:
which XX is your version of SQL Server. So XX would be 10, for SQL Server 2008R2, could be 11 for SQL Server 2012 and so on. The difference between the local msmdsrv.exe file located in your Power BI Desktop\bin folder with the other one you can find under your SQL Server installation folder is that the one which Power BI Desktop runs is a console programme while the other one is a Windows service programme.
How to find Power BI Desktop local port?
There are various methods you can obtain the port number. In this post I explain three of them.
Finding Power BI Desktop local port using Windows Command Prompt (CMD)
Using DAX Studio
Finding local port number from Power BI Desktop temp directory
Another cool feature is adding Synonyms to the model. Power BI Synonyms can significantly improve the Q&A and query experience. With synonyms, we can now add descriptions to the data model objects such as tables, columns and measures in the Power BI Desktop. The descriptive information could include names that the end-users may possibly use to refer to an object or abbreviations used across the business. Addin these descriptions or, as the name suggests, synonyms makes using Q&A even easier for our customers to find what they are looking for. The customers don’t know all table, column or measure names. Defining a standard list of names for tables, columns, or measures makes Q&A much more helpful.
For instance, we can add the following synonyms:
Note:The following tables and columns are from AdventureWorksDW.
Original Name
Object Type
Synonym
FactInternetSales
Table
Internet Sales, InternetSales
OrderQuantity
Column
Order Quantity, Order Qty, ord qty
SalesAmount
Column
Sales Amount, Sales Amt, Internet Sales Amount, Internet Sales Amt
TaxAmt
Column
Tax Amount, Tax Amt
Freight
Column
freight
OrderDate
Column
order date
How it works
It’s easy to set up synonyms in Power BI Desktop. Switch to Model view, then click “Synonyms” from the “Modeling” tab from the ribbon. Then, simply enter the synonyms.
After we publish a Power BI Desktop model to Power BI Service, the synonyms will play a great role in Q&A so that when the customer types “ord qty” the Q&A engine will recognise it as “OrderQuantity” and display the results. It’s really cool, isn’t it?
But let’s think a little bit out of the box. What if we add some translations as synonyms? Hmm. I think it would be really great if a Spanish customer could type Spanish column names in Q&A rather than English. I added some translations to FactInternetSales columns and DimDate columns.
Thanks to Google Translate for French and Spanish translations. Sorry French and Spanish guys, if the translation looks funny.
Now, I publish the model to Power BI Service. To do so, just click on “Publish” from the “Home” tab from the ribbon.
As I promised in my earlier post, in this article I show you how to leverage your Power BI Desktop model using Query Parameters on top of SQL Server 2016 Dynamic Data Masking (DDM). I also explain very briefly how to enable DDM on DimCustomer table from AdventureWorksDW2016CTP3 database. We will then create a Power BI Desktop model with Query Parameters on top of DimCustomer table. You will also learn how to create a Power BI Template so that you can use it in the future for deployment.
In the previous post I explained how to create dynamic data sources using Query Parameters. You also learnt how to use Query Parameters in Filter Rows. In this post you learn :
Using Query Parameters on top of SQL Server Dynamic Data Masking (DDM)
I’m not going to provide much details about DDM as you can find lots of information here. But, to make you a bit familiar with Dynamic Data Masking I explain it very briefly.
Dynamic Data Masking (DDM)
Dynamic Data Masking (DDM) is a new feature available in SQL Server 2016 and also Azure SQL Database. DDM is basically a way to prevent sensitive data to be exposed to non-privileged users. It is a data protection feature which hides sensitive data in the result set of a query. You can easily enable DDM on an existing table or enable it on a new table you’re creating. Suppose you have two groups of users in your retail database. Sales Persons and Sales Managers. You have a table of customers which in this post it is DimCustomer from AdventureWorksDW2016CTP3. This table contains sensitive data like customers’ email addresses, phone numbers and their residential adders. Based on your company policy, the members of Sales Persons group should NOT be able to see sensitive data, but, they should be able to all other data. On the other hand the members of Sales Managers group can see all customers’ data. To prevent Sales Persons to see sensitive data you can enable Dynamic Data Masking on the sensitive columns on DimCustomer table. In that case when a sales person queries the table he/she will see masked data. For instance he see uXXX@XXX.com rather than user@domain.com.
Create a table with DDM on some columns
It’s easy, just put “MASKED WITH (FUNCTION = ‘Mask_Function’)” in column definition. So it should look like this:
CREATE TABLE Table_Name (ID int IDENTITY PRIMARY KEY, Masked_Column1 varchar(100) MASKED WITH (FUNCTION = ‘Mask_Function’), Masked_Column2 varchar(100) MASKED WITH (FUNCTION = ‘Mask_Function’),
…
)
GO
Alter an existing table and enable DDM on desired columns
As you guessed you have to use “ALTER TABLE” then “ALTER COLUMN”. Your T-SQL should look like:
ALTER TABLE Table_Name ALTER COLUMN Column_Name1 ADD MASKED WITH (FUNCTION = ‘Mask_Function’);
GO
ALTER TABLE Table_Name
ALTER COLUMN Column_Name2 ADD MASKED WITH (FUNCTION = ‘Mask_Function’);
A template is basically a Power BI file that represents an instance of a predefined Power BI Desktop which includes all definitions of the Data Model, Reports, Queries and parameters, but, not includes any data. Creating Power BI Templates is a great way to ease the deployment of existing models. Creating templates is very easy, you just click File –> Export –> Power BI Template. We will look at this more in details through this article.
Scenario
You are asked to implement a new level of security on customers’ data (DimCustomer on AdventureWorksDW2016CTP3 database) so that just privileged users can see the customers’ email, phone numbers and residential address. Privileged users are all members of “SalesManager” database role. You are also asked to prevent “SalesPerson” database role to see sensitive data. But, all members of both “SalesManager” and “SalesPerson” database roles can query DimCustomer table. The users should NOT have SQL Server logins.
