peterfrb wrote The problem is that when I use the Browser to drag and drop a customer's name as a column heading, I do not get a full compliment of names.  In order for me to see all the available names, I must first select the "Show Empty Cells" option.  The peculiar thing about this is that nothing else -- no other column, no row, no measure, no filter -- but the one column is present.  In other words, nothing that I can see should be classifying these names as "empty", since nothing else is present with which to compare it.

Actually, there is something to compare. A cube always has a default measure. If you issue something like SELECT NON EMPTY Customers.MEMBERS ON 0 FROM <cube> that is using the default measure. So, if the default measure is empty, the customer won't show up.

Also, note that the browser in BIDS puts NON EMPTY on by default. When you choose the Show Empty Cells option, the MDX is simplified by removing the NON EMPTY keywords.

peterfrb wrote Acting on a previous recommendation, I ran the SQL Server Profiler, added my column to the browser, and used the Profiler's trace component to capture the MDX query I generated.  I then ran SQL Server Management Studio  in "Analysis Services" mode, created a new MDX query, and copied and pasted the traced view into the MDX query.  However, I always get errors when attempting to parse the queries I've imported.  I'm so new to this whole process, I don't know how to diagnose or evaluate the errors I'm getting.

The problem is how the browser in BIDS operates. It creates temporary sets to hold certain information (the list of customers in this case.) So the query is actually two queries: one that builds the set, and then one that uses the set. Why does BIDS do this? I'm not really sure. But the only way to go is to start writing the MDX queries by hand.