SQL Server SSAS (Multidimensional MDX) - an Introduction

It used to be that we had to install a cut-down version of SQL Server. Instead, let's now install a version with the full functionality of the Enterprise edition - for personal use only, though.

Now let's go through the process of installing SQL Server.  I'll also go through the various editions of SQL Server (e.g. 2008, 2012).

Now that's the back engine has been installed, we've got to install SSMS (SQL Server Management Studio).

We are going to use two versions of AdventureWorks - the first we will use to create our first cube, and the second will be for a practice activity. Let's download them now.

Let's have a look at AdventureWorks, using SQL Server Management Studios, and have a look at the tables used.

Every cube needs to be creating within a project. Let's create our first project.

Now that we've got a project, let's have a look around at the various panes. Also, just as T-SQL has its own terminology (such as "window functions"), SSAS has its own words. Let's get used to facts, measures, dimensions and cubes.

Now that we've got a project, we need to connect to some data. Let's create a data connection with our AdventureWorks database.

Just like a painter needs colors, a cube designer needs tables. Let's add some tables into a Data Source View.

We've got a data source and a dimension - let's now create a cube, and pull a couple of tables into the cube.

So on my computer, I find that I can't see the cube that I've just created. Let's add a role for my account, and change the impersonation for the data source, and see whether that solves the problem. Then let's take a quick look at the cube.

Now we've got a cube, let's look at it. First of all we'll open Excel from Visual Studio directly. Then we'll go to Excel 2016 and we'll add a database connection, and create a PivotTable, and see how that is different from traditional PivotTables. We'll do the same again in Excel 2013 to see how that it is different.

Let's add an extra table and create a dimension from this table.

Now we've created a dimension, let's re-process the cube, and then view it in Excel again.

Not only can you use T-SQL in SSMS, you can also use MDX, creating SELECT statements from your cube. But be warned - the syntax may look like T-SQL, but it is actually quite different.

What else can we do with cubes? Why don't we create a report as well, using SQL Server Reporting Services.

Now it's your turn. Let's create a second cube, this one using the AdventureWorksDW file. We'll go through what you have got to do - and then, over to you.

As a reminder, here are your instructions.

Let's run through, and create this second cube.

Existing dimensions can be modified, by adding additional attributes. But what if one of the attributes you want to add it merely a translation of an existing one? Instead of addition translation attributes, let's translate an attribute instead.

Now that we've got the cube working, let's add another table into the DSV, and then after trying to create a new dimension, we'll replace this table with a query.

Hierarchies can be very useful in organising data, such as dates into years, months and days, and cities into countries, states and cities. Let's create a 2-level hierarchy from our date dimension.

Now let's change our hierarchy from 2 levels (year - date) to 3 levels (year - month - date). It's now at easy as it sounds. Let's find out what problems we encounter.

• prevmember
• nextmember
• lag
• lead
• head
• tail
• cousin