MDX Function and Operator Reference - Professional SQL Server Analysis Services 2005 with MDX - page 1053

These functions calculate the statistical covariance across x-y pairs of
values. The y numeric expression is evaluated over the set to get the y
values for each pair. If the x numeric expression is present , then it is
evaluated over the set. Otherwise, the cells formed by the set are evalu-
ated within the current context, and their values are used as the x values.
The biased population formula for covariance is as follows:
CrossJoin ( set1 , set2 ) Returns: set Standard (additional behavior in
AS2005)
set1 ⊛ set2 Returns: set Extension: AS2005, AS2000
These functions return a set forming the Cartesian product of the two
sets (except for a Microsoft extension to the semantics noted below). The
two sets must represent different dimensions; you will get an error if the
same dimension appears in both of them. CrossJoin () only takes two
sets as arguments. However, because it takes two sets as input and re-
turns a set as its output, you may nest multiple calls to CrossJoin () to
take the Cartesian product of three or more dimensions. Following the
same rules used for composing tuples by hand, the order of the dimen-
sions in the resulting tuples is the same as the order of dimensions in the
set arguments. Using an asterisk between two sets, as with {set1 ⊛
set2}, is a Microsoft-specific synonym for CrossJoin (). The expression
set1 ⊛ set2 ⊛ set3 is the same as CrossJoin (set1, CrossJoin (set2,
set3)).
In Analysis Services 2005, if the two sets are composed of tuples from
the same base dimension but different attribute hierarchy-dimensions,
then only the combinations of tuples that actually exist in the underlying
dimension are returned. There is no way to produce tuples that do not
have corresponding entries in underlying tables.
See also: Extract (), Generate (), Distinct (), NonEmptyCrossJoin ()
Set. Current Returns: tuple Standard