Category Archives: SSAS

The Power BI dashboard in public preview – available also without Office #powerbi

Microsoft released a new version of Power BI in preview mode, including many new visualizations that are immediately available to all existing subscribers also in production, such as the long waited treemap, combo charts (combining line chart and column chart), and more. These features are available only in HTML5 visualizations, so you can only use the new features online. Microsoft shown these visualizations several times this year (PASS BA Conference in San Jose, and PASS Summit in Seattle), so now this is finally available to anyone. But there is much more!

Power BI Dashboard is a new service, now in public preview (unfortunately only in United States, not sure about which other countries are supported by now, certainly not Europe), that does not require an Office 365 subscription and, more important, provide a design experience on desktop also without having Excel or Office at all. In other words, there is a separate Microsoft Power BI Designer that enables you to:

  • Import data with Power Query
  • Create relationships between tables
  • Create data visualizations with Power View (running the latest HTML5 version locally in a desktop application)

This very first release does not include the full data modeling experience we are used to in Power Pivot, so you cannot create calculated columns or measures, but hopefully this will come in the next updates. In this way, you can use Power BI with a separate “data model” environment that is not tied to Excel. You can have an older version of Excel, or no Excel at all, and still design your data model with the Designer.

The goal of this app by now is to simply offer an offline design experience, and I have to say that performance of data visualization is very good. With the Designer you design data models and reports. Once published in the Power BI web site, you can “consume” data, but you can also modify the report and “pin” objects to a dashboard, so that you can build your own custom dashboard, such as the Retail Analysis Sample you can see below.


You can create datasets getting data from several SaaS applications, such as Dynamics CRM, Salesforce, GitHub, ZenDesk, SendGrid, and Marketo. You can also connect to live Analysis Services through a new gateway named Power BI Analysis Services Connector and use new native mobile apps for Power BI. Support for iPad should be already available (again, depending on countries, it seems not available in Europe by now). Future support for iPhone and Windows tablets has been already announced.

This is a very interesting evolution of the Power BI platform and I look forward to use it with real data and real users! Many tutorial videos are available on YouTube.

Rename table, columns, and measures in #ssas #tabular without breaking existing #dax

A feature that many people require in SSAS Tabular is the ability to refactor existing names, doing a correspondent rename in all existing objects in the model. I agree that this is an important feature that should be added in the development environment, but this will help only the development cycle. Once you release a Tabular model, the names you published becomes part of queries created by the users. For example, if you save a pivot table, the objects selected (table, column, and measure names) are all part of the MDX code that is generated automatically by Excel. If you rename something… at the next refresh, Excel will remove renamed objects from the Pivot Table. In less politically correct way, any renaming operation potentially break existing reports.

Some years ago I heard from a student in a course that they were using translations in SSAS Multidimensional to avoid this issue. They were developing using English, but since users were using other languages (I was in North Europe) they had a decoupling layer between internal model names (in English) and external ones. Any rename operation was completely painless in this way.

A few days ago, I reminded that and I thought if it was usable for Tabular… and I discovered that you can also use a translation for the same primary language of your model! This is really interesting and deserve to be investigated more. Please, read my article Frictionless Renaming in Tabular Models with Analysis Services and give me your feedback if you have time to test this approach. I am curious to see possible issues of this technique. Thanks!

DAX Studio 2.0 finally available! #dax #ssas #tabular #powerpivot

Darren Gosbell announced on his blog that DAX Studio 2.0 is available. This new release has a completely new user interface, a renewed architecture and also work as a standalone executable without Excel. Yes, you read it right. You can use DAX Studio on a server without having to run it in Excel. You can now connect with Remote Desktop to your SSAS Tabular server and create DAX queries with a decent editor.

I don’t see any good reason to use SSMS to run DAX queries now. Until yesterday, the zoom in the editor was the only one, but now also this feature is available in DAX Studio. I could spend other lines of this blog post describing the other new features, but I don’t see how this could be more important than doing the only thing you should do at this point: download, install and use it!

Power Query now connects to Analysis Services #ssas #mdx #dax

The last update of Power Query finally supports Analysis Services Multidimensional and Tabular. I waited this version for a very long time, but now it’s finally here!

Chris Webb already wrote an interesting blog post with several useful link and informations. 

You can connect to both Multidimensional and Tabular, but queries are generated in MDX. For this reason, I consider Multidimensional as a primary source by now. Many improvement can be done for multidimensional, whereas Tabular would benefit from DAX support at first.

I want to share my feedback and I already look forward to future improvements: please use the comment to this blog post to share your priorities for these features I would like to see.

Expose Keys

Each Hierarchy in Multidimensional has one or more levels, corresponding to dimension attributes.
Each attribute has up to three properties:

  • Key
    • Can be multipart, which means it has one or more columns, each column can have a different data type
  • Name
    • This is always a string. If not defined, it corresponds to the Key, which must have a single column. If the attribute has multipart key, the Name has to be defined in the Multidimensional model
  • Value
    • This property is optional and is not used often. However, it could represent the name in a numeric value and it’s used when the key cannot be used to represent the corresponding value. This property is not shown to the user but can be used in MDX calculations.

The Name is the one shown in the user interface of power Query. However, when you extract data from a cube, most of the times you need the key column(s) in order to create relationships with other query. For example, a very common scenario is creating three tables, corresponding to two dimensions and a fact table. The keys used to connect the dimension tables to the fact table are mandatory, but are not exposed in the attribute name visible in the user interface of a cube.
Thus, exposing the Key is very important. Please remember that the Key can be a multipart key, so you might have one or more columns to add.
If a user request the Key, by default I would show all the columns of a multipart key so he can remove those columns that are not required (but usually you’ll need all of them). Since the Value is not requested so often, I would expose it as a separate menu.

Surrogate Keys and Other Hidden Attributes

Depending on the cube design, it would be nice exposing those attributes that are hidden in the cube.
For example, in a well-designed cube, the model should not expose the surrogate keys in a visible attribute, because this would create a dependency in reports that would break a filter based on a surrogate key in case the relational tables are reprocessed and surrogate keys are regenerated. The general idea is that a surrogate key does not have a semantic meaning. Thus, it shouldn’t be exposed to an end user.
However, if you are importing several tables from a Multidimensional model, trying to create one table per dimension and one table per fact table (maybe by changing the cardinality, reducing the number of dimensions used), then you should import the surrogate keys too, at least for SCD type 2. Unfortunately, there are no info in the cube model that could help you discriminating between SCD1 and SCD2, so I’m not sure about what could be the best practice in this case. Probably, hidden attributes should be accessible only by advanced users, exposing them by default could be really confusing and I would avoid that.
This is an area where it’s hard to make a right choice, a compromise is required.

Related Properties

An attribute can have other properties related to it. By default, all browsable parent attributes can be considered attribute’s properties. However, when we talk about attribute’s properties we usually refer to the non-browsable attribute. Each non-browsable attribute is just another attribute. Non-browsable attributes are not shown in the list of attributes of a dimension, but they should be available to the user that want to import data for a certain column. The user interface could be designed in several ways for that:

  • Show attribute’s properties in a window from which the user can drag&drop – but maybe confusing – the UI should show only properties existing for a particular attribute and each attribute might have different properties. User interface might be a concern here.
  • Automatically import all the properties of an attribute (maybe by asking confirmation to the end user?) when adding that attribute to a query. Then the user can remove the columns that are not required in the Query.

Show Formatted Measures

Sometime it might be useful to import the formatted value of a measure. However, I would not replace the number imported today with the formatted value, because the latter could be a string that does not have any semantic meaning. Providing the option of importing the formatted measure as an additional column in the PowerQuery window would be very welcome, but don’t just replace one with the other.

Invisible attributes

Currently, Power Query shows all the attributes of a dimension, which in general is a good thing. However, I would put a checkbox that shows/hides invisible attributes. By default, I would show only visible attributes, because this is what user would be more familiar with. THe “show invisible attributes/columns” should be an advanced view.

Multiple selection without measures

if you select attributes from several dimensions without selecting a measure, you obtain as a result the Crossjoin between the tables you selected. In my opinion, this is counterintuitive and useless: I cannot imagine a use case where this would be meaningful. A much better solution would be importing every dimension as a single table, just as you do when you select many tables from SQL Server. It is the user that will handle joins between table, if necessary. My suggestion is to keep the existing behavior (import a single table) only when you import also a measure, even if I would like to be able to import all the dimensions and the set of measures as separate tables in the data model – creating one query for each dimension and one query for each measure group (or for each cube – not sure about the better approach here).

What are useful tools and resources for DAX developers? #dax #powerpivot #tabular

At the last PASS Summit I received an interesting question: is there a list of all the useful (I would say necessary…) tools for DAX developers? My answer was… “no, but this is an interesting topic for a blog post”.

In the meantime, I thought that a page to keep updated would have been better, and of course an easy-to-remember URL is equally important. So here is the URL where you will find an updated list of tools and resources useful to any DAX developer:

Of course, feedback are welcome!

Don’t use SUMMARIZE to sum your data–or just be careful #dax #ssas #powerpivot

During the last PASS Summit I and Alberto Ferrari had long discussions at SQL Clinic with some of the developers of our loved tools. Sometime you really have to dig in the internals of certain feature to understand why there are some “unexpected” behaviors for certain functions. One of the discussions was about SUMMARIZE. This function can be very powerful (after all, it’s a way to do a join between related tables…) but also very dangerous, because of the way it implements its logic (especially for the ROLLUP condition).

The rule of thumb, that we already mentioned in the past, is to use SUMMARIZE only as a way to execute a sort of SELECT DISTINCT, and not to create column to aggregate values. Use ADDCOLUMNS for this other job. We previously mentioned mainly performance reasons for that, but now we have a more complete description of why you should avoid SUMMARIZE for computing aggregations: you might obtain a different result than the one expected. The complete discussion of the issue and of the workarounds is included in the new article All the secrets of SUMMARIZE written by Alberto Ferrari.

BI Announcements at PASS Summit 2014 #sqlpass #powerbi #powerpivot

This morning the PASS Summit 2014 started in Seattle and during the keynote there was several announcements from Microsoft. I’m considering here only the ones about Business Intelligence (you will find other blogs around about SQL Server).

  • In the coming months, Azure SQL Database will get new features such as column-store indexes, which can be very interesting for creating data marts on the cloud
  • Another upcoming feature in SQL Server will be an updateable columns-store index on in-memory tables. Real-time analytics will like this feature.
  • For a store analysis, an interesting demo using Kinect capturing heatmap to display which areas of a shop store have been visited more using Power Map. Just a demo, but it’s an interesting idea and the best big data demo I’ve been so far (something you can implement in the real world using big data technologies without being Twitter or Facebook).
  • New Power BI dashboards: many new visualizations and a new user interface to place data visualizations on a dashboard (similar to the grid you have in DataZen if you know that product)
    • You can connect to your data source from the cloud, without creating a local data model and sending it to the cloud
    • Q&A is integrated in the new user interface – the web site is a domain, it seems not in SharePoint
    • Q&A generates the report in HTML5, no Silverlight signs here
    • The entire editing is done in a web browser – a preview of that was presented at PASS BA Analytics keynote, this seems a more refined version (still not available, however)
    • TreeMap is available as a new visualizations
    • You can upload an Excel file from your disk or from OneDrive – just Excel file, no Power Pivot data model required (it is created on the fly on the cloud?)
    • Combo chart combining line and bar chart visualization available
    • Private preview now, public preview available soon
    • Request access to public preview on
  • Azure ML is publicly available for free in trial mode

The Power BI story seems the real big news. Combining this with the fact that you can query *existing* on-prem databases on Analysis Services without moving them on the cloud opens up interesting scenarios. Many questions now about when it will be available and how it will be deployed. Interesting times ahead.

Power Query support for Analysis Services (MDX)

Today at TechEd Europe 2014 Miguel Llopis made the first public show of Power Query support for Analysis Services.

This is still not available, but it should be released soon (hopefully it will be our Christmas gift!).

Here is a list of features shown:

  • It should be able to query both Multidimensional and Tabular
  • Generates query in MDX (no DAX by now)
  • Load one table at a time (but a query can mix dimensions and measures)
  • Shows dimensions, measures, hierarchies and attributes in Navigator
  • Use the typical Power Query transformations working on a “table” result
  • You import one table at a time

I think the last point deserves an explanation. When you write a query in Power Query, the result is a single table. If I want to build a Power Pivot data model getting data from an existing cube in Analysis Services, but with a different granularity, I have to run one query for each dimension and one query for the fact table. Depending on the definition of the cube, this could be easier or harder, because original columns could have been hidden because measures are exposed instead. Moreover, the result of a measure that is not aggregated with a sum (imagine just an average) could be impossible to aggregate in Power Pivot in the right way.

Thus, if you want your user to take advantage of Power Query, make sure you expose in a model measures that can be aggregated to compute non-additive calculations (such as an average!)

Now I look forward for receiving this Christmas gift!

A strange behavior of AutoExist for MDX on Tabular models #powerpivot #ssas #tabular

Alberto Ferrari wrote an interesting article about a strange behavior of AutoExist in normalized data models. I always say that a star schema is the best practice in Power Pivot and Tabular data modeling. The issue described by Alberto is another good reason to avoid snowflake schemas.

I think that an example is better than many words. Consider this simple measure working in a star schema where all product attributes (such as Category and Subcategory) are in the same denormalized DimProduct table:

SalesOfBikes := CALCULATE ( [Sales], DimProduct[Category] = "Bikes" )

If you have a snowflake schema with DimProduct, DimProductSubcategory and DimProductCategory tables, you have to write a much longer and complex DAX formula in order to obtain the same result:

SalesOfBikes :=
    DimProductCategory[EnglishProductCategoryName] = "Bikes",
        FILTER (
            ALL ( DimProductSubcategory ),
            IF (
                ISFILTERED ( DimProductSubcategory[EnglishProductSubcategoryName] ),
                CONTAINS (
                    VALUES ( DimProductSubcategory ),
                    DimProductSubcategory[ProductSubcategoryKey], DimProductSubcategory[ProductSubcategoryKey]

Which seems crazy, and actually it is…

The reasons are interesting and well described in the AutoExist and Normalization article on SQLBI.

Questions and Answers about SSAS Tabular Models #ssas #tabular

I recently delivered the online session Create your first SSAS Tabular Model at 24HOP Pass Summit, which recording is now available here. I received many questions and did not have enough time, so I answer now in this blog post.

How do you prevent a user from aggregating certain measures where the result would be invalid (example: unit margin %)?
In Tabular you do not have the notion of “aggregation”. Every measure evaluates a DAX expression in a particular filter context. Imagine this as a SELECT statement in SQL that consider only the rows filtered by a WHERE condition. You might “remove” the visualization of a value in a measure by using DAX functions such as ISFILTERED, ISCROSSFILTERED and HASONEVALUE. Since you do not have the ability to “intersect” expressions in different dimensions as you can do in MDX using Tool dimensions, you do not have the problem of invalidate certain combinations (such as “Unit” and “Margin %”.

How do you incrementally refresh this new Tabular model that was created?
You can add data to an existing partition, create new partitions or reprocess existing partitions.

Can you connect to SSRS report from Power View to drill down to the actual data rows?
Currently this is not available. In general, you cannot add link to external URLs in the current version Power View.

Can you create those views within the model instead of in the database, like you can with cube creation?
No, in Tabular you do not have the notion of “Data Source View” like the one you have in Multidimensional. By the way, I do not consider a “best practice” embedding queries in a DSV in Multidimensional. If you can create views on the relational database, you simplify troubleshooting in case of data quality issues reported by end users (any DBA can check the views used, even without any knowledge about BI development tools).

Can many to many relationships work in tabular model?
In this version, you cannot create a many-to-many relationship directly in the data model, but you can apply many-to-many relationships in DAX formulas. The good news is that performance are usually faster than equivalent models in Multidimensional.

Any changes for Tabular model between 2012 and 2014 versions of SSAS?
No, there are no changes in Analysis Services between 2012 and 2014. SQL Server 2014 is a release that added new features only to the relational engine of SQL Server.

Can you give a few examples of benefits over using multi-dimensional cubes?
Tabular is easier to use, it is usually faster and it requires no maintenance (Multidimensional requires maintenance of aggregations as data volume grows and data distribution changes). Multidimensional has features not available in Tabular (e.g. custom rollup formulas, MDX Script, dynamic formatting for measures).

How about the role playing dimensions concept can work in Tabular model? Is there an equivalent of role-playing dimensions in Multidimensional in Tabular?
Role-playing dimensions are a usually a bad idea in Multidimensional, because you cannot rename hierarchies, attributes and member names. This result in confusing pivot table when you browse the data, considering that the only difference is the dimension name, which is not directly visible in the pivot table itself.
Tabular does not support role-playing dimensions, but you can overcome that limitation by using DAX, enabling inactive relationships for specific calculations (you can define multiple relationships between the same tables in Tabular). However, if you want to offer navigation in different role-playing dimensions to the user, the best practice for both Multidimensional and Tabular is to import the same table multiple times, renaming data and metadata.

Can you install regular SSAS and the Tabular model on the same server?
Yes, you can install several instances of SSAS on the same server. You run the setup multiple times and choose for each instance whether it has to run as Multidimensional or Tabular. You make this choice during the setup.

Can you use a server based Tabular model for O365 Power BI sheets? (on premise data – SharePoint cloud spreadsheets)
Not yet – at the moment (September 2014) you can only publish a Power Pivot workbook and then refresh it getting on-premise data through the Data Management Gateway.

Can I create a Tabular Model using a query as the source?
Yes, but remember that this is not a best practice. Creating SSAS Tabular or Multidimensional models, the best practice is getting data from SQL views, without modifying the query on SSAS side. In this way, the content of a Tabular table will match the result of a view in SQL. This simplifies the maintenance and the support operation. Any DBA can check the result of a SQL view without having to open an Analysis Services project just to figure out where actual data comes from.

How I can manage access of the data by level (For example: first group have access to 2014 year, but second group have access to all periods)?
Role-based security allows you to create row-level security for each table in the data model. Thus, you can filter the rows of the tables you want to hide to a certain group of user. You define filter conditions using logical DAX expressions that are evaluated when each user creates a connection to SSAS Tabular model.

Does the language M has something to do with the SQL server or it is used only in Excel?
The language “M” is used only by Power Query, which is used only in Excel and in Data Management Gateway today. The “M” language can produce transformations in SQL, but you cannot obtain an “M” version of a SQL query.

What happens if the source data of a Tabular model is not a star schema?
A star schema is the best data model for Tabular. However, you can have more complex data models, but keep in mind that more tables and relationships might cause slower performances at query time.

How do you handle multiple dates in a fact table that you want to attach to a date table?
This is a question similar to the role-dimension one. You can import the Date table multiple times (possibly by renaming columns and content, reflecting the “role” of the dimension in these names), or you can create multiple relationships between Date dimension and Fact table, activating one relationship for each measure through the USERELATIONSHIP function in a CALCULATE statement.