Skip to Content
0
Former Member
May 19, 2010 at 01:28 PM

BPS Layouts - aggregates

37 Views

Hi everybody,

in my consolidation web application I use a lot of different BPS layouts. Almost all layouts base on multiarea which contains of two basis areas.

Layouts consists 4 different columns which get data from 2 different cubes:

Column1 Column2 Column3 Column4

Cube1 Cube2 Cube1 Cube1

Columns 2,3 and 4 are comparison columns. Which data are show in columns 2,3, and 4 depends on the version (e. g. 250, 260, 750) which can be chosen in drop down boxes. This drop down boxes are connected with BPS variables.

Recently we have performace problems with some layouts. I check our aggregates and I have found something strange.

I will write an simple example what I have found out:

We have small aggregates for all versions: 250, 260 and for 750. Additionally we have a big aggregate with all version.

Example 1

Column2 Column3 Column4

Version 250 250 250

Data are read from aggregate 250 - correct

Example 2

Column2 Column3 Column4

Version 260 260 260

Data are read from aggregate 260 - correct

Example 2

Column2 Column3 Column4

Version 250 260 250

Data are read from aggregate with all versions and not from 2 different small aggregates 250 and 260. If I read the data from big aggregate with all version performance is of course bed. Can you explain me, why it is so? I have done already a lot of different tests and it is always like that.