Datapoint Value Accuracy

Post Reply
User avatar
Steve Vincent
Site Admin
Posts: 1054
Joined: Mon May 12, 2008 8:33 am
OLAP Product: TM1
Version: 10.2.2 FP1
Excel Version: 2010
Location: UK

Datapoint Value Accuracy

Post by Steve Vincent »

Not sure if you'd call this a bug or a feature. I have a cube where a consolidated value is a whole number but its children get split in decimals, down to about 6dp, to get it spot on. I needed to change the consolidation keeping the proportional spread of its children the same, so i used data spreading to do that and it all worked fine.

As i had about 15 to do, and 2 sets of years for each, i decided to copy the values and paste them over all the others. The cube view was formatted to 0dp and so i was surprised when i recalculated it to see the end value was 1768 and not 1770 as required.

Code: Select all

1,770.000000	136.233061	136.147245	170.184056	136.147245	136.147245	170.184056	136.147245	136.147245	170.184056	136.147245	136.147245	170.184056

Code: Select all

1,768	136	136	170	136	136	170	136	136	170	136	136	170
I know what its doing, i'm just bemused as to why. When it copies the data all it picks up is the formatted value, not the underlying full value. Thats different to how Excel works where even if the data is formatted to 0dp, the value stored is the original 6dp or whatever it was.

Is that the same in later versions of TM1, and is that how others feel it should work? To me that's wrong, it should be the same as Excel especially where it has such high integration between the 2. Its one of those things that you get so used to it working one way that you never give it 2 thoughts in a different system, and in the end just causes more work.
If this were a dictatorship, it would be a heck of a lot easier, just so long as I'm the dictator.
Production: Planning Analytics 64 bit 2.0.5, Windows 2016 Server. Excel 2016, IE11 for t'internet
User avatar
jim wood
Site Admin
Posts: 3958
Joined: Wed May 14, 2008 1:51 pm
OLAP Product: TM1
Version: PA 2.0.7
Excel Version: Office 365
Location: 37 East 18th Street New York
Contact:

Re: Datapoint Value Accuracy

Post by jim wood »

For some reason this rings a bell. Isn't there a cfg setting relating to exactly this issue? I would search the forum just in case. (It might be me going mad mind)
Struggling through the quagmire of life to reach the other side of who knows where.
Shop at Amazon
Jimbo PC Builds on YouTube
OS: Mac OS 11 PA Version: 2.0.7
Herman Moller
Posts: 70
Joined: Thu May 22, 2008 3:38 pm

Re: Datapoint Value Accuracy

Post by Herman Moller »

Steve,

Examine the following parameter for the tm1s.cfg:

SpreadingPrecision : Floating point arithmetic on computers is not 100% precise. When a computer calculates very small numbers, a margin of error is applied to the calculation. If the computer adds a set of numbers, and the resulting sum is close to the target value within the margin of error, the sum is considered accurate.The margin of error for TM1 spreading operations is controlled through the SpreadingPrecision parameter. The default value is SpreadingPrecision=1e-8. You can adjust the precision as required for you application.
Post Reply