Page 1 of 1
CFG Parameter 'ForceReevaluationOfFeedersForFedCellsOnDataChange'
Posted: Wed Jul 20, 2016 12:02 pm
by sumitgoyal2290
Hi,
I am new to TM1 and currently looking at certain features of TM1.
Can anyone help me with the cfg file parameter 'ForceReevaluationOfFeedersForFedCellsOnDataChange'.
I have tried to test this in my personnel machine by creating a conditional feeder in order to fed the cell. But I have to restart the server or run TI cube process feeder in order to refresh the feeder.
I am not able to see any change by using the parameter ForceReevaluationOfFeedersForFedCellsOnDataChange.
Working on Tm1 version 10.2
Thanks
Sumit
Re: CFG Parameter 'ForceReevaluationOfFeedersForFedCellsOnDataChange'
Posted: Wed Jul 20, 2016 4:21 pm
by Steve Rowe
Hi Sumit,
Please post details of your test system so that we can explain why it may not be working.
It is more likely that your test is not constructed correctly than there is a bug in TM1.
I could be wrong but my understanding is that attribute value changes do not trigger re-evaluation.
If you are actually requiring this to be live in a proper TM1 system then I would suggest you look hard at your design. It will make the performance of the system worse, IMO any design that requires this cfg to in place has significant issues.
Cheers,
Re: CFG Parameter 'ForceReevaluationOfFeedersForFedCellsOnDataChange'
Posted: Wed Jul 20, 2016 7:09 pm
by tomok
Steve Rowe wrote:I could be wrong but my understanding is that attribute value changes do not trigger re-evaluation.
I believe the issue is that string value changes don't trigger re-evaluation. Granted, most attributes are strings but they don't have to be.
Re: CFG Parameter 'ForceReevaluationOfFeedersForFedCellsOnDataChange'
Posted: Wed Jul 20, 2016 8:45 pm
by lotsaram
tomok wrote:I believe the issue is that string value changes don't trigger re-evaluation. Granted, most attributes are strings but they don't have to be.
Actually it's the opposite. String cells don't require a special parameter to force feeder re-evaluation, never have. The default bahaviour of string cells unlike numeic cells is that feeders will re-fire on each value change, unlike numeric cells which only refire when changing from null to a non-zero value. We have used this functionality many times, it can be particularly useful in a picklist driven application where a picklist value encodes an element name which can then be used in a feeder. (try it I promise it works).
Where the OP may be getting mixed up and I think what Steve might have been getting at is the situation ..
['my cell'] => DB('some cube', !dim1, !dim2, AttrS('dim3', !dim3, 'lookup'), 'some measure');
When the value of the "lookup" attribute changes the end result is that the correct cell(s) is not fed but some other cell is still fed. Even when ForceReevaluationOfFeedersForFedCellsOnDataChange is used this won't make a shred of difference as the feeder is only going to re-evaluate when the (leaf) values defined by "my cell" change, not when the attribute values in dim3 change. The only way to deal with this situation is with a TI using CubeProcessFeeders (or deleting feeders and restarting the server whihc is more drastic but which will also clear out the old overfed cells.)
Re: CFG Parameter 'ForceReevaluationOfFeedersForFedCellsOnDataChange'
Posted: Thu Jul 21, 2016 12:24 pm
by tomok
lotsaram wrote:Where the OP may be getting mixed up and I think what Steve might have been getting at is the situation ..
['my cell'] => DB('some cube', !dim1, !dim2, AttrS('dim3', !dim3, 'lookup'), 'some measure');
When the value of the "lookup" attribute changes the end result is that the correct cell(s) is not fed but some other cell is still fed. Even when ForceReevaluationOfFeedersForFedCellsOnDataChange is used this won't make a shred of difference as the feeder is only going to re-evaluate when the (leaf) values defined by "my cell" change, not when the attribute values in dim3 change. The only way to deal with this situation is with a TI using CubeProcessFeeders (or deleting feeders and restarting the server whihc is more drastic but which will also clear out the old overfed cells.)
That was what I was talking about, when you use a string value as a "lookup" type feeder. Once the original mapped cell has been fed, any change to the lookup value will not cause the newly mapped cell to be fed. To be honest, that's the only situation where I've ever used a string to feed something.
Re: CFG Parameter 'ForceReevaluationOfFeedersForFedCellsOnDataChange'
Posted: Fri Jul 22, 2016 11:12 am
by sumitgoyal2290
Hi
Thanks everyone for updates.
My scenario is that I have 3 cubes: cube 1, cube 2 and cube 3. Data in cube 3 is flowing from cube 2 based on some condition in cube 1.
If certain condition is met in cube 1 then only data should be fed in cube 3 from cube 2.
Now I am assuming that 'ForceReevaluationOfFeedersForFedCellsOnDataChange' parameter should work here. But my issue is that unless I run CubeProcessFeeders TI(Or restart the server), numeric data is not getting fed.
Re: CFG Parameter 'ForceReevaluationOfFeedersForFedCellsOnDataChange'
Posted: Fri Jul 22, 2016 11:56 am
by Steve Rowe
Please post your actual code.