I'm running an MDX query against a MOLAP cube built on 99% performance.
When I run the query which produces a very large cellset (too big for
the MDX Sample Application) the app hangs with a CPU usage of 100%. Is
there anything I can do about this ? A way for me to optimise
performance to speed up the query (although as I've left it 'running'
for more than 1/2 hour I'm beginning to think its overheating rather
than chugging on)

I'm currently testing the MDX to the limit of what the user may query
and I have used a lot of crossjoins and a lot of .Members. Children
doesn't seem to make a difference but would Descendents() maybe be
better?

Hi Ben,

What for MDX query do you use? It depends a lot on calculated memember in
your query.
What target do you pursue? You get huge cellset, so what do you do with it?
What usefulness of such cellsets?

Thanks,
Vladimir Chtepa

"Ben" <gringogordo (AT) hotmail (DOT) com> schrieb im Newsbeitrag
news:1136547462.732638.82320 (AT) g49g2000cwa (DOT) googlegroups.com...

The Cube has been created to allow the way various financial sets are
spent over time according to various criteria. The MDX then has to
apply these criteria to the Cube and bring back a cellset.

In practice the users will choose specific criteria that will produce a
fairly small cellset. Also I may be able to reduce the size by
aggregating the totals. At the moment due to the fact that I'm quite
unfamiliar with MDX I am getting a cellset back and processing it in
code (as I am struggling to get the set back that I want). Although
the solution probably lies in the MDX it concerns me that even when
setting a timeout on the command it seems to crash when the cellset is
very large.

The Cube has been created to allow the way various financial sets are
spent over time according to various criteria. The MDX then has to
apply these criteria to the Cube and bring back a cellset.

In practice the users will choose specific criteria that will produce a
fairly small cellset. Also I may be able to reduce the size by
aggregating the totals. At the moment due to the fact that I'm quite
unfamiliar with MDX I am getting a cellset back and processing it in
code (as I am struggling to get the set back that I want). Although
the solution probably lies in the MDX it concerns me that even when
setting a timeout on the command it seems to crash when the cellset is
very large.