TM1 Forum

Hi
I'm wondering whether I should think this is as expected or it is a bug:
I have a feeder from cube A to B, having a nested DB on the right hand side of the feeder (to better understand: I have customers in the source cube, customers and suppliers in the target; on the RHS of the feeder, a nested DB queries the assigned supplier for each customer from a lookup cube so that one customer from the source feeds only one customer and one supplier in the target)

I have created a new sandbox and changed the customer-supplier assignment in the sandbox only, and then refired feeders by saving the rule in the source cube. It seems that the feeder did not work in the sandbox; the rule is calculating correctly, showing values on the new supplier on N: level but they are not fed.

this is 10.2.2

is this as expected with sandboxes?

thanks
matyas

I have never attempted to alter a feeder in a sandbox only. I would be HIGHLY surprised if this worked.

another reason why I'm not a big fan of sandboxes

btw I did not attempt to change a feeder; just the nested DB on the RHS of the feeder returns different value for the sandbox than for the base for one single cell; but maybe this is how you understood I attempted to change a feeder for the sandbox only

That's an interesting one. I would think that sandboxing is primarily designed & engineered for pure data consolidation type applications (although obviously it also works with rules) and so I wouldn't be surprised to find that if a string mapping cell that fired a "conditional" feeder based on the the string value mapping to an element name for example didn't fire when changed in sandbox mode. Although I wouldn't be surprised I would still regard this as a bug and if you have an example I think it would be worth going through the effort and annoyance of building a test case to convince IBM of the bug.

I'm a fan of "driver based" planning models utilizing picklists where the picked drivers often will have an effect on feeding. I would want this to work in sandbox mode. I haven't seen an example of this NOT working but also I haven't thoroughly tested such a scenario. Maybe you are doing something slightly more complicated and have found a hole.

Hi
My concern is mainly that Contributor also works in a kind of sandboxing mode (or private workspace mode?) by default and this can cause problems there, too.

Thanks lotsaram and tomok for your input!