STET rule behaviour

[mvaspal]

Hi

I would gladly hear your views about the following behaviour of the STET rule. I could also simply reproduce it in the GO_NEW_STORES sample by creating CubeA and CubeB as 2-dim cubes from Month and PriceandCost.

Rules in CubeA:

Code: Select all

['Jan' , 'Unit cost'] = N:STET;
['Unit cost'] = N: DB(CubeB, !Month , !Price and Cost);

Then the steps done:
1. Enter data into CubeA, jan, unit cost

2. Commenting out the first line -> unit cost for jan is displayed from CubeB as expected. As much as I know, the formerly entered data for jan and unit cost should still reside in the cube "behind the scenes".

3.
a) Commenting the second line out, too -> data comes back as stored in CubeA
b) Leaving the second line as it is and enabling the first line again -> data lost, zeroes displayed.

I could find this thread but I think this is a different case from that.
http://www.tm1forum.com/viewtopic.php?f=3&t=483&p=2948

Thanks
Matyas

[Duncan P]

You haven't said what version you are on but if it's 10 then it might be something to do with RulesOverwriteCellsOnLoad.

[mvaspal]

I'm sorry, I first saw this on 9.5.2 and then I reproduced it on 10.1
I tried adding RulesOverwriteCellsOnLoad=F on 10.1 now but still no luck.

[lotsaram]

The title if your post is misleading as the issue you're bringing up is nothing to do with STET per se but the apparent change in behaviour of rules deleting or not deleting cube data present before the rule was applied. It is true that the behaviour you were expecting (comment out the rule, data previously input to the cube reappears) is what "traditionally" has been the case. But as this feature was undocumented I guess IBM did not feel obliged to inform when this changed somewhere around 9.4. or 9.5 to rules having the effect of deleting underlying data should there be any. Maybe enough people have complained or raised concerns as in 10.1 there does seem to be the new config parameter referred to by Duncan that would seem to be intended to toggle this behaviour. I never tried it using it in practice but I assume from reading the documentation that the intention is to be able to restore the former behaviour.

[mvaspal]

Hi

The title if your post is misleading as the issue you're bringing up is nothing to do with STET

I would argue with this because of the differences in 3a and 3b. If I remove the rule from the second line (3a), then data comes back as I expected. For example, if I have only one line of rule, ['Unit cost'] = N: DB(CubeB, !Month , !Price and Cost); then it does not seem to cause the same issue as with STET.

I also agree that this new 10.1 parameter seems to addess similar issues but as my quick sample shows, it has not changed the behaviour of my example in 3b.