[Mondrian] Bug in getNamedSet(List<Id.Segment> nameParts) ?

Pedro Alves pedro at neraka.no-ip.org
Thu Feb 19 13:40:41 EST 2009

On Thu, Feb 19, 2009 at 10:28:00AM -0800, Julian Hyde wrote:
> There seems to be a contradiction in my rules of evaluation: I stated that
> when a set is evaluated, it is evaluated using the slicer as context, but
> not context from the axes. But this particular set is used during the
> evaluation of the slicer, so there's a cyclic dependency. Should this give
> an error? Or should it use the default member of the dimension?

I admit I was expecting it to behave like if the defaultMember was there;
but if an error was thrown, at least it would have saved me a couple of
hours debugging the code.

> Pedro, So that the time we have spent on this issue does not go to waste,
> can you do experiments on SSAS2005. Write some minimal queries to nail down
> its exact behavior with regard to the context that sets are evaluated in.
> Log a bug if mondrian is not consistent with SSAS2005.

Unfortunately, I don't have access to a SSAS2005, but Matt Campbell
reported the exact same behaviour in 2000 and 2005, so if this is an issue,
mondrian is not the only one facing it


Pedro Alves

More information about the Mondrian mailing list