Hi,
I am new to MSAS and DTS and would like direction in the following
design.

I have designed a cube based on my database and would like to create a
scheduled DTS task to update the cube.

My database is in effect a logging database of call events based on a
datetime field eg. "recordtime". I use this table as the fact table. I
also have another identity column on this fact table eg. "sequenceid".

What I want to do is daily incrementally update the cube based on the
last time I incrementally processed the cube.

So - Can I store the "Run parameters" ie. max(recordtime) or
max(sequenceid) whenever I update a cube so that my filter expression
is based on these parameters.ie.
SELECT * from table where recordtime > "RUNPARAMETER.MAXRECORDTIME" OR
sequenceid > "RUNPARAMETER.MAXSEQUENCEID"

thanks for your help
bill

Sounds like a viable solution - a couple of things to consider along the
way:

- If the fact table is large, it should preferably have an index on the
filtered field, to reduce query times

- Make sure that either cube and "Run parameters" are both updated, or
that none of them is (can use a transaction)


- Deepak

Deepak Puri
Microsoft MVP - SQL Server

*** Sent via Developersdex http://www.developersdex.com ***

Deepak,

Thanks for your reply.

Yes the fact table is large (actually a horizontally partitioned view)
and is indexed.Actually performing other specific aggregation using
stored procs and works well.

But in this case - I need to publish analysis data hence my question.
Now - where can I store the Run Parameters ?
Excuse my lack of MSAS knowledge but - can I store these parameters
within the cube ? If so - how do I update them on successful completion
of Incr processing of the cube.


I must be missing something simple here ...

bill

Deepak Puri wrote: