[Mondrian] Why Mondrian rank value is reversed from SSAS2005

Julian Hyde jhyde at pentaho.com
Thu May 29 18:47:20 EDT 2008


> I did some comparisons on the behavior of Rank() between 
> Mondrian and SASS2005 (reported in
> http://pub.eigenbase.org/wiki/MondrianRankFunction.)
> 
> It seems that Mondrian choose to return reversed rank value 
> from SASS (the first point in the above link). JVS guessed 
> the argument would be "it's more natural to give rank 1 to 
> the highest value" given the trick posted in 
> http://www.sqljunkies.com/WebLog/mosha/archive/2006/03/14/mdx_
> ranking.aspx. I would appreciate if someone could confirm 
> that or point out any other benefits of this choice.

If our behavior is different than SSAS that's a bug. In particular I didn't
realise we were getting RANK wrong. Please log a bug (or bugs).

Julian 





More information about the Mondrian mailing list