Building Basic Dimensions and Cubes - Expert Cube Development with SSAS Multidimensional Models - page 60

• Processing objects in the wrong order : For example, if you process a

dimension, then update the underlying relational data so that there are new
rows in the dimension table and new rows related to them in the fact table,
and then process your cube without reprocessing the dimension again first,
you'll again run into key errors.

• MDX Script errors : It's very often the case that when you make structural

changes to a cube or a dimension, these changes break MDX calculations on
the cube. For example, you might rename a dimension, but still have MDX
code where the old dimension name is referenced. When a cube is processed,
the last thing that happens is that all of the code on the MDX Script is
executed and if it now contains syntax errors, the whole processing operation
will fail. This is extremely frustrating when it happens, and it can lead to a
lot of wasted time. If you are making a lot of structural changes, it can be a
good idea to comment out the whole of the MDX Script before you do any
processing, and only uncomment it when you have successfully processed
your cube.

Summary
With processing complete, you can take a look at your cube for the first time, either
in the Browser tab of the Cube Editor or in your client tool of choice. Now is a good
time to reflect on what we've seen of the cube development process so far. We've
created a very basic cube from a single fact table and a few dimensions rather than
attempting to build something more complex. This has allowed us to get a feel for
our data and have something to show our users quickly so that they can check if
we're on the right track. We then built a single Data Source and Data Source View.
Since we spent time getting our data modeling right earlier, there was very little to
do here other than connect to our data warehouse and select the tables or views we
want to work with. Next, we built a few of the less complex dimensions we need,
configuring attribute relationships and creating user hierarchies as necessary. Finally,
we ran the New Cube wizard to build our basic cube, then deployed and processed it
so that it can be queried.

In the next chapter, we'll go on to look at how we can deal with more complex
problems when designing dimensions.