Hi all,
First, a bit of background: In our cubes, all dimensions have the surrogate key defined as the dimension key and we typically only update or insert rows in our data warehouse dimension tables. Recently we needed to reload the data warehouse due to a problem
with the server, and this has had the effect of breaking existing pivot table reports. In looking at the underlying MDX, the surrogate key value is referenced in the queries, but because the dimension tables were reloaded, the surrogate keys have changed and
the sid references are now invalid, or incorrect.
Is there a recommended way to fix this? Everything I've read indicates that the sid should be the key column for a cube dimension. Is this not correct?
Luckily we do not yet have heavy usage of the cubes, but that fact that a reload appears to break all reports makes me very nervous. Have I missed something fundamental in our cube design?
That's a few questions, but I guess the one that's paramount is: how do I fix this?
Thanks,
-Mark