I want to create an extract file from a Cube ... but I want to use the field
code / key in the extract not the description / name.

Looking at Foodmart I want something like the following:

Select

{ [Measures].[Unit Sales] }

on columns,

--non empty

crossjoin

(

{

descendants( [Customers].&[Canada] )

},

{

descendants( [Product].&[Beer] , , leaves )

}

)

on rows

FROM Sales

This gives the long names in the output, .keys is little use as the levels
are not known in advance, and I don't want both name and key by using
DIMENSION PROPERTY before the "on rows" - the file would be too big.

Any help appreciated.