Date Dimension script optimized for Analysis Services KeyColumn property (Part 1)

When attribute dimensions are built, there is a choice to design KeyColumn property as a Composite Key (consisting of several key columns) or a Simple Key (consisting of only one key column). To ensure proper data rollup in the user-designed hierarchies, sometimes (and especially in the Date/Time dimension) it is necessary to use Composite Key. For example, Quarter, Month and Week levels need to use Year key along with the original (Quarter, Month or Week) key to be placed properly under their parents. Usually, Composite Keys take more space (by duplicating the same data) in the Analysis Services data storage structures and generally less optimized than Simple Keys. Even additional internal data structures can appear in the situation when String type is used as a Key Column (very bad practice), thus making key structures far from optimal for data loading and navigation. The last statement relates to the Simple Keys as well.