[Mondrian] RE: Degenerate dimensions

Julian Hyde jhyde at pentaho.com
Sat Mar 14 22:40:39 EDT 2009

My physical schema change could/should handle this. I guess I should write
up a description of what new-style mondrian schemas will look like. But note
that in SSAS 2005 there is a matrix of measure groups vs. dimensions, and
for each dimension there are several ways to join the dimension to the
measure group's fact table, including regular (foreign key from fact table
to dimension table), via intersection table, and the dimension table is
identical to the fact table. Linkages in the new mondrian schema will be
modeled similarly.
Physical schema change is long overdue, and is next on my backlog after 3.1
is out, so I won't consider doing it any other way.


From: Matthew.Campbell at thomsonreuters.com
[mailto:Matthew.Campbell at thomsonreuters.com] 
Sent: Friday, March 13, 2009 10:02 AM
To: wgorman at pentaho.org; julianhydepentaho at speakeasy.net

Julian and Will,


We're just about to start our next phase of development, and one of the
upcoming features is enabling degenerate dims in our model.  Our degenerate
dims are common across multiple fact tables, however, and we need a way to
use them across tables.  For example, I might want to report on "Claim ID",
even though that is a degenerate dimension on multiple claim tables.


There's an enhancement request for something like this already:
35302&atid=414616> &aid=1802271&group_id=35302&atid=414616


Could either of you hazard a guess at how difficult this change would be?
We could probably pick this up if you think it's a straightforward change.





PS:  I have been distracted from investigating enhancements to native
evaluation for the past month with the completion of our release milestone.
I hope to pick up analysis again soon and will likely have questions.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20090314/5d038a80/attachment.html 

More information about the Mondrian mailing list