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

John Sichi jsichi at gmail.com
Fri Feb 20 19:44:57 EST 2009


On Thu, Feb 19, 2009 at 1:10 PM, Julian Hyde <jhyde at pentaho.com> wrote:
> Thanks for taking this on, Matt.
>
> This stuff should be extra test cases in SsasCompatibilityTest (attached,
> because it is a work in progress, not yet checked in).
>
> A few behaviors seem reasonable when a calc member in the slicer uses a set
> that is dependent on the calc member: give a cyclicity error, use the
> default member, use the all member, use the calc member. The experiment
> should give a clear answer which of these SSAS2005 does. Also, if the set is
> subsequently used outside of the slicer, does it give the same result? Am I
> even correct in asserting that calc sets use the slicer context?

Rushan and I investigate some related behaviors a while back; she has
published the writeup here:

http://pub.eigenbase.org/wiki/MondrianQueryEvaluationContext

Nothing conclusive but may be useful background info.

JVS




More information about the Mondrian mailing list