TM1 Forum

It appears that Feeders in a rulessheet will only be triggered (processed) when in the same rulessheet the Skipcheck is actived. If not, defined (external) Feeders will not work.

Yesterday I ran into a situation where I had a cube where the Skipcheck statement was hashed out, but the Feeders statement was still in place. The Feeders section contained (correct) feeder statements to another cube, but apparantly these are not feeding. If I activate the Skipcheck the external feeders are working correctly (but then the numbers aren't correct - I guess that's why Skipcheck was hashed out). I'm currently revising the internal feeders of the cube and expect that everything will be okay after this. But I thought it was worth mentioning this behaviour here.

Michel

If skipcheck is removed or commented out then the need for feeders is removed as all calculated values will have a flag on it to let the system know they are there. Th eproblem is that the flag will be in place regardless of whether data is there. Memory usage potentially goes up and supress zeros will not work. Switching skipcheck off removes all flags. You therefore need feeders to put relevant flags in place.

Having skipcheck on is the only reason for feeders, thats why they don't work work.

Hi Jim,

I know how feeders work . The cube with the hashed out Skipcheck statement does not need the Feeders (well, it does from a performance perspective, but that's beside the issue here). This cube is calculating correctly without using Skipcheck/Feeders. The key thing for me was that if I have a cube without Skipcheck I can never feed to another cube from this cube. If I want to create an external feeder I need to implement Skipcheck on the source cube first.

Michel

Apologies for teaching you to suck eggs. I'm not sure your first email was clear about the external cube bit. Nevermind. Anyway I now know what you mean and yes you are right you can't feed to another cube without switching skipcheck on. A bit of a pain. Have you thought of creating say a feeder cube? Take the data from the first cube via TI and load it into a another. Then create the feeders in this cube with skipcheck switched on. ot the nicest of solutions but it is one I guess,

Jim.

No worries. I was confronted with an old model that needs to be revised thoroughly. I was going for a quick win but found I had to go all the way implementing S&F bottom-up in the cascade of interlinked cubes. As writing S&F on all cubes was one of the things on my list it will be my first action now.

Michel

Good luck.

Hi,

sorry to raise a very old post from the dead but I have 2 questions:

1) Is it still the case that feeding to another cube without skipcheck will result in no feeding?
2) Is there any scenario where having skipcheck off but feeders on makes any sense?

Maren

MarenC wrote: ↑Mon Jul 20, 2020 10:27 am Hi,

sorry to raise a very old post from the dead but I have 2 questions:

1) Is it still the case that feeding to another cube without skipcheck will result in no feeding?
2) Is there any scenario where having skipcheck off but feeders on makes any sense?

Maren

I don't think that #1 has ever been the case. You can have Feeders without SkipCheck.

Whether it makes sense it harder to say. Plenty of times I have Feeders in a cube that otherwise doesn't have rules. You could have a case of a cube with rules but no SkipCheck (say the cube had ony 2 or 3 dimensions so sparsity isn't an issue and so there's no need for feeders) and thi scube did have feeders to some other cube that did contain SkipCheck (becasue maybe that cube has > 3 dimensions and so sparse consilidation is required.)