Had anyone had any experience in using views as sources for all their cubes
and dimensions? Any thoughts and inputs would be appreciated.

Thanks

We have a production system running off views for over a year, no
problems experienced - and the views are helpful as a means to abstract
from changes in the table structure.

Regards,
RuiDC

Hi Rui,

Can I ask is there anything special about the views you are using? Are they
indexed views or views with schema bindings?
Are you also using them in both dimensions and cubes?

The reason I asked is that after a couple of months into our development, we
keep on encountering Fact Table = 0 errors. This happens often and we're a
bit tired of going into the cubes and editing the fact count property.

Any inputs would be appreciated.

Thanks,
Joseph

"RuiDC" <rui.dacosta (AT) ubs (DOT) com> wrote

I've not come across that problem. Is hould mention however, that in
our case, the views are on top of a sybase database using ODBC drivers.

It sounds like you're getting some form of corruption, have you applied
SP4 for AS2K? I know we had some unrelated corruption problems that
have gone away since we installed post SP3 hotfixes and rebuilding
(some of the fixes only work as prevention to corruption, they cannot
fix it after it has occurred).

The other thing to try is to clone your problem cubes - by doing a copy
paste - although you would then need to reprocess - and replace the
originals.

But either way, i don't think the issue would be caused by you basing
on views, as there should be no difference do AS2K given that it just
runs a query against the RDBMS anyway - ie. what could be different
between a query against a table and a query against a view? AS2K as a
client, should be agnostic as to the source.

Maybe someone else can prove this logic wrong...