TM1 Forum

Hi Tm1 Experts,

May I just ask this issue. I'm showing the previous month's comments. After I put my feeders, when I try to save the rule file, it hangs.
Here is my rule. Is there something wrong with my feeders?

There are other measures here. All of them referenced from other cubes.

FEEDSTRINGS;

SKIPCHECK;

['Prev Month Comment']=S:
DB('CubeA',!Dim1,ATTRS('Reference Period',!Period Month,'Prev Month'),!Dim3,!Dim4,'Comment');


FEEDERS;

['Comment']=>
DB('CubeA',!Dim1,ATTRS('Reference Period',!Period Month,'Next Month'),!Dim3,!Dim4,'Prev Month Comment');

Thanks in advance!


Bunch

Guys,

I am able to save the rule file now. But when I try to suppress, it's missing. When not suppressed, the text's there.

Guys sorry, it doesn't work yet. It just saved well because I removed the S from the FEEDSTRINGS; made it FEEDSTRING;

Still have the problem.

Is it because it treats blanks cells as string also? A non-empty string? Just a theory.

Hello Bunch,

To be able to feed a string cell, you must use the syntax 'FEEDSTRINGS' not 'FEEDSTRING'. Given that your session hangs when you use the correct syntax, it may well be that the feeders only trigger when the correct syntax is used. I suspect that your 'hang' is actually TM1 processing a large amount of feeders. There is no indication form your post about the size of the cube or the other dimensions contained within it. That would be a critical consideration.

What you might like to do is to devlare a more succinct feeder on the left hand side (LHS) of your feeder statement. for example;

['Comment','Version1','Product1','Cost centre1']=>
DB('CubeA',!Dim1,ATTRS('Reference Period',!Period Month,'Next Month'),!Dim3,!Dim4,'Prev Month Comment');

Start with what you know will generate the smallest amount of feeders and build up from there. As you do this, look at the memory and CPU on task manager on the TM1 server. If you see memory rising considerably, or CPU flatlining, then you know you are just stressing the system by processing a large amount of feeders. Keep doing this, until you are using a simge cell to feed a single cell, if you must.

You can then test if your feeders have worked for the slice that is defined in the LHS of your feeder statement. If sucessful, begin to drop dimensions off the LHS until you have sufficient feeders for what you require and not to stress the server.

Hope this helps.

-Chris

Unfortuantely Sir, you are correct. I have an overwhelming eleven dimensions. And the things is, initially, all cells are subject for input, so I can't be sure on the intesections that can be identified as my feeders.

What I think is to run a process to flag. This is a numeric flag. Anyway, this report is just used monthly. In that way, I have the time to run the TI.

How do you find this idea Sir?

I cant imagine that you need to capture commentary and previous month commentary for every instersection of every cell. There may well be some merit in housing the commentary in one cube and the other data in another.

An alternative would be to run a TI process to copy the commentary accross. It would not be live, but I somewhat suspect that you do not to see last months commentary appear live in the current month. You could put it on a chore to run once a month.

Hope this helps.

-Chris

An alternative would be to run a TI process to copy the commentary accross. It would not be live, but I somewhat suspect that you do not to see last monhs commentary appear live in the current month. You could put it on a chore to run once a month.

Yes Chris. This is more preferable. What's the point of running a TI to load flag values as feeders. That will just add physical data and memory usage when the flags feed.

This is right.

Thanks for the solution.