Bad performance when export data from rule-derived cells

[phamtinkt88]

Hi all,

I'm trying to export data from cube to .csv but it took a long time to finish (30 mins).
I use TI to export data with CubeView datasource. Below is the view :

• Currency - 1 element
  Version - 1 element
  Year - 1 element
  Month - 12 elements
  Store - 30 elements
  SKU - 45000 elements
  Measure - 1 element (Rule-derived element)

When i export data from the view with non rule-derived Measure element, TI performs very good !
So i knew that bad performance was caused by rule-derived element.
Are there any ways to optimize TI that export data from rule-derived element ?

Thanks for your help !
Tin

[Duncan P]

There could be a number of things at work here.

I am presuming that you have selected the option in the TI source definition to skip zero values. If not then you should.

Then you will also need to have SKIPCHECK as the first statement of the rule and you will need a feeder in the FEEDERS section to define which instances of Measure have non-zero values. Otherwise the engine will need to run the calculation for all of the measure cells in order to know which produce non-zero results.

If all of this is already in place then it would probably be helpful to see the rule statement for Measure to see how complex it is and whether you would be better off doing it in the TI script.

[phamtinkt88]

There could be a number of things at work here.

I am presuming that you have selected the option in the TI source definition to skip zero values. If not then you should.

Then you will also need to have SKIPCHECK as the first statement of the rule and you will need a feeder in the FEEDERS section to define which instances of Measure have non-zero values. Otherwise the engine will need to run the calculation for all of the measure cells in order to know which produce non-zero results.

If all of this is already in place then it would probably be helpful to see the rule statement for Measure to see how complex it is and whether you would be better off doing it in the TI script.

Hi Duncan,

skip zero was checked.
And I've already add SKIPCHECK & FEEDERS, but due to my calculation in rule quite complex.
There are 8 levels in SKU hierrachy, user first input at level 8 (Consolidate cell) then allocate amount to level 7, complete level 7 then allocate level 6 ... process finish after they complete allocating level 0 (Leaf) => rule apply for each single level, both consolidated and leaf. So i don't think SKIPCHECK & FEEDERS will help in this case.

[Duncan P]

So what is the rule? Is there any way it could be more efficient to implement in the TI?