Are conditional scripts possible using MDX - ssas

Question

In SQL the following structure is possible
IF <condition>
BEGIN
<do something>
END;
Can I use conditions in a similar way within MDX ?
EDIT
Specifically I'd like to somehow test if a set exists and use that as the condition in the IF statement.
IF TheFollowingReturnsRow(
select *
from $SYSTEM.MDSCHEMA_SETS
where SET_NAME = 'Set_Custom' and
CUBE_NAME = 'MyCube') THEN
<do something>
END IF

Answer 1

Depending on what you want to achieve, it is definitely possible to use IF statements in MDX:
http://technet.microsoft.com/en-us/library/ms145517.aspx
SCOPE ([Customer].[Customer Geography].[Country].MEMBERS);
IF Measures.CurrentMember IS [Measures].[Internet Sales Amount] THEN this = 10 END IF;
END SCOPE;
Other alternatives include the IIF(Logical_Expression, Expression1, Expression2) and the CASE WHEN ... THEN ... ELSE ... END struct.

Answer 2

I think MDX is mainly aimed to be a query language, and hence the SELECT statement is the most evolved statement. There are some other DML statements like UPDATE CUBE, DRILLTHROUGH, or CALL, but except for UPDATE, they all just read data, and are used less frequently.
Then, there are some DDL commands, but these are limited to slight adaptations of the existing cube or your current query session, you cannot create new cubes or dimensions using MDX. The "real DDL commands" for SSAS that create cubes or change dimensions etc. are XMLA commands, not MDX statements.
DCL statements like GRANT or REVOKE are not available in MDX.
And for the fourth category of statements that you find in SQL, the procedural statements like if statements (as opposed to if expressions) or looping constructs, as well as server side cursors, MDX does not offer any support as well. These statements are the part of SQL that is not very standardized between different DBMSes like Oracle, SQL Server, DB2, mySQL, to name just a few, while basic DML and DDL is more or less standardized - and to some extent all major relational databases support ANSI SQL. Even for relational databases, this is often a later add-on used mainly in stored procedures, triggers (as a special case of procedures), and batches. MySQL was widely used for many years without having support for these statements which were only introduced in version 5.
For stored procedures, Analysis Services relies on COM (deprecated) and .net, where you have full blown procedural features, and just makes its object model available to .net or COM. Anyway, stored procedure support is limited to functions returning one object (which may be as complex as an MDX set containing a lot of data), or procedures returning result sets. The latter cannot interact with other MDX statements, but just be called with CALL procname(parameters). And for batches, MDX just relies on the hosting environment that calls the batch to do all the conditional branching and looping.
And in fact, MDX has a much more restricted use than SQL, so to some extent, it is just a "domain specific language", that is very good for some purposes, but has many weaknesses in areas that it is not intended for.

Answer 3

I would suggest moving all such conditional logic to cube calculation script if possible.
There it is possible to use scope statements which are much more efficient.
Cube calculation script is stored at the server side SSAS. You can have calculations and some other logic defined. These will be available to MDX queries. For example you can create new members and also new sets. Also it is possible to override the values loaded/calculated if needed. Here is a simple explanation on how it works http://www.sqlservergeeks.com/blogs/AmitK/sql-server-bi/65/implementing-calculations-in-ssas-using-mdx-part1
If I use the example given by Dan. You can even do things like
SCOPE ([Customer].[Customer Geography].[Country].MEMBERS);
SCOPE ( [Measures].[Internet Sales Amount]);
this = 10;
END SCOPE;
END SCOPE;
If conditions are very slow, but after such a conversion, they should work at a reasonable speed.