Apache OpenOffice (AOO) Bugzilla – Issue 23737
Iterations calculates wrong result
Last modified: 2004-01-27 12:28:05 UTC
I have a business application calculating selling prices for certain items using iteration. I have been using oleo so far, but now thought I would transfer to scalc. But the iterations doesn't work any more as in oleo. I narrowed down the problem to a very small example but I couldn't find a place to insert this example on this page. The dangerous thing is that the result is clearly wrong.
Created attachment 12038 [details] Try changing the transport cost here and check the selling price...
Hi Peter, please have a look. Frank
Hi, after a recalculation (SHIFT+CTRL+F9) the results are as expected. Adjusting menu->tools->options...->spreadsheet->culcalte->iterative references->minimun change to 0.01 also seems to resolve the issue. Best regards, Peter
No, changing the minimum change to 0.01 doesn't make it work! It still needs the recalc command repeated (many times) until the value settles down to the correct one. When iterations are on, the recalc should be automatic and really do the calculation until nothing changes more that the minimum change. At least here that doesn't happen. The iteration process stops too early.
> changing the minimum change to 0.01 doesn't make it work! How many steps are configured?
First tried with 1000 (default) then with 1000000. Still doesn't work. If you do the change I mentioned together with the example you notice that changing the transport costs only will not change the sell price at all, until you press recalc, and even then you have to press recalc many times before you get tha right result. If this doesn't do the same for you, could it have something to do with the locale used? I am using a locale where the decimal point is a "," ??
Well, when I open the attached bugdoc I also need two recalcs to ,get all the expected results. Changing the settings in tools->options does not eliminate all problems on windows 2000. @Eike: What do you think about this issue? Is it maybe duplicate to SO internal ID: 86392? Peter
Peter, This is not related to #86392# at all. But it is the same effect as in the second part of issue 956, the minimum change test isn't applied to all formula cells involved in the circular calculation, but only to the injecting formula cell. Btw: would have been helpful if the description stated what exactly is considered wrong in the calculation, namely the 0.02 divergence of the interim values in cells I15:I16, as the overall calculation shows no error. Resolving as dupe of issue 956. Eike *** This issue has been marked as a duplicate of 956 ***
Close dupe.