Hi:

I have a cube that updates every hour. There is one partition for each
hour's data. There is only one fact table.

I found out that if I feed the cube each hour with the fact table containing
same data (no change to the contents of the table), the aggregation for each
hour is only 5 seconds.

But if I feed the cube with different data (same fact table with NEW set of
data, number of rows also changes), the aggregation for each hour took 5
minutes. Any one can think of why?

Thanks

A couple of things to consider.

1. If you feed the partition the same exact data it is likely that the data
is already cached in memory on the RDBMS. So your read phase might be much
faster.

2. It could be that the aggregates are larger on the other partition, which
would require more processing time.

Are you using the same method to process both partitions? If you've already
processed a partition once, there are other processing options available
(like reaggregate for example) that would make such a comparison unfair.

The best thing to do to determine the difference is to break the processing
down into its phases. Try processing with 0 aggregates and see what
happens. This will give you the time AS needs to read the data and write
the fact table. Then add aggregates and select the reaggregate option.
That will give you how much time is devoted to aggregatation. In your case,
you could probably be able to spot the difference by just watching the
processing in Analysis Manager.

Sean


--
Sean Boon
SQL Server BI Product Unit

--
This posting is provided "AS IS" with no warranties, and confers no rights.
Use of included script samples are subject to the terms specified at
http://www.microsoft.com/info/cpyright.htm.



"Polaris" <etpolaris (AT) hotmail (DOT) com> wrote