use multiple aggregates on infoprovider parallel

Question: Help please!

I have an infoprovider with a lot of data.
There are KF with exception aggregation. Some have reference characteristic 'Day' and some 'Contract'.
To speed up performance I woulkd like to make aggregates, but
there is not much flexibilty there because I need both Contract and Day
in the aggregate to serve the exception aggregation.
It would be nice if I can create 2 aggregates that are accessed in parallel. One for each reference characteristic.

Example:
aggregate 1 contract and day level
3000 * 1000 = 3000000 records in aggregate

aggregate 2 contract level only
3000 records in aggregate

aggregate 3 day level only
1000 records in aggregate

Anybody know how to realize this?
It seems impossible to let the query acces both aggregates and combine them.

Thanks in advance.

Answer:
The engine cannot access multiple aggregates at the same time for one info-provider but is in the pipe line for development (most generic DW provide this function). However, as work around you can create copies of your cube but change the structure of the cube to the desired level, then create a multi-provider over you new cubes. A multi-provider will then execute the queries against the under-lying cubes in parallel.

Kevin

Answer:
Thx Kevin for your quick reply.
This is also the only solution I had in mind.
I will try this to see if it works as expected.

I do not really like the idee that I have to store all cube data twice...
but as a temporary workaround it might be ok.

You have any idee when this functionality will be available?

Rgds,
Ronald

Answer:
Hi Kevin,

I assumed that by changing the structure you mean putting keyfigures
with reference Day in one cube and the ones with reference COntract in the other.
The aggregation level of both cubes are the same.

I quickly tried the solution above, but it does not work as expected.
The query on the multicube is indeed split up into 2 parallel subqueries, but using RSRT I see that both of the subqueries have the same
level of aggregation. Both have the reference characteristic Day and
Contract in it.
I expected one to have the Day and one to have the Contract, because
one KF with reference Day is in one cube and one KF with reference Contract is in another cube.

I have a feeling this problem cannot be solved untill this new functionality
is available.
Copyright ?2007 - 2008 www.jt77.com