Hi to all and be merciful with beginners...

I have a cube where fact tables contains member values NOT appearing in
dimension keys. I wish I had a warning of this when processing my cube, but
I haven't: cube process completes successfully, and I have to check data by
confronting grand totals calculated by hand, then adding missing dimension
keys. Where am I wrong? I gambled uselessly with every option in the Process
Cube Settings, no erroro is raised, no log file written....
First time ever that I'm begging for an error!
Can anybody help me?

AS gets data from your star schema by issuing a query that joins the fact
table with all the dimension tables. These joins will eliminate all the fact
data that does not contain valid dimension keys.

Does this make sense?

Regards
Jamie Thomson
http://www.conchango.com

P.S. There is one situation where this (i.e. joining with the dimension
tables) doesn't happen. i.e. When you have used the Optimise Schema
functionality. Forget about this for now though!


"Moco" wrote:

Jamie,
Thanks for prompt response. You are totally right. Is there a way to force
AS to raise an error instead of simply "eliminating" fact records? That is,
can I impose referential integrity somehow? I tried to run Optimize, and
received the message that my dimensions (all of them) could not be
optimised.
thx



"Jamie Thomson" <jamie.thomson (AT) donotspam (DOT) conchango.com> ha scritto nel
messaggio news:8E71F278-9E7E-4F90-A7A6-1AFC051B93BC (AT) microsoft (DOT) com...

[Forget about "Optimise Schema". I shouldn't have mentioned that - it only
complicates the matter!]

The answer is no, you can't get it to force an error. The only place to
enforce referential integrity is in the underlying database.

Regards
Jamie Thomson
http://www.conchango.com


"Moco" wrote: