Welcome, Guest. Please login or register.
July 18, 2025, 08:48:34 AM

Login with username, password and session length

Search:     Advanced search
we're back, baby
*
Home Help Search Login Register
f13.net  |  f13.net General Forums  |  General Discussion  |  Topic: Math is hard - serious bug in Excel 2007 0 Members and 1 Guest are viewing this topic.
Pages: [1] Go Down Print
Author Topic: Math is hard - serious bug in Excel 2007  (Read 3188 times)
Trippy
Administrator
Posts: 23657


on: September 24, 2007, 09:25:39 PM

Just in case some of you are using Excel 2007: Bug in Excel 2007

Basically some calculations that should come out to 65535 are showing as 100000.

Edit: fix:

http://blogs.msdn.com/excel/archive/2007/10/09/calculation-issue-update-fix-available.aspx
« Last Edit: October 10, 2007, 06:08:54 AM by Trippy »
Venkman
Terracotta Army
Posts: 11536


Reply #1 on: September 26, 2007, 06:44:38 AM

Jesus. I just tried that on three earlier versions of Excel I have and it calculates fine. How do you mess something like that up?! It's almost like they had to try...
Roac
Terracotta Army
Posts: 3338


Reply #2 on: September 26, 2007, 07:51:50 AM

Jesus. I just tried that on three earlier versions of Excel I have and it calculates fine. How do you mess something like that up?! It's almost like they had to try...

Because of the binary representation.  2^16=65536, and 2^20 is 100,000 in hex.  2^16 is also two bytes.  Somewhere something broke, had an overflow, whatever.

-Roac
King of Ravens

"Young people who pretend to be wise to the ways of the world are mostly just cynics. Cynicism masquerades as wisdom, but it is the farthest thing from it. Because cynics don't learn anything. Because cynicism is a self-imposed blindness, a rejection of the world because we are afraid it will hurt us or disappoint us." -SC
Xanthippe
Terracotta Army
Posts: 4779


Reply #3 on: September 26, 2007, 10:07:06 AM

Computer Science 101 ftw.
Morat20
Terracotta Army
Posts: 18529


Reply #4 on: September 26, 2007, 12:47:55 PM

Computer Science 101 ftw.
Overflow errors are so freakin' common. We try to right overflow-proof classes as much as we can here, but you always miss something.
Chenghiz
Terracotta Army
Posts: 868


Reply #5 on: September 27, 2007, 11:41:00 AM

I'd like to take this opportunity to express the fact that Office 2007 is a badly designed crock of shit.
Lt.Dan
Terracotta Army
Posts: 758


Reply #6 on: September 27, 2007, 04:04:42 PM

I thought about this and can't really see how this would break any of my spreadsheets. 

Cell referencing? Errr, 65k rows and you should be working in a real database.

External database calls?  See above.

Dates? Maybe.

Code? Maybe, but what the hell would you use integers of 65k or higher for?  An array that big would break excel.

Finance calculations?  We work to four decimal places in unit prices - the likelihood of hitting 65535.00 is basically nil.

Now if it had been floating point calc I'd be more worried...
sinij
Terracotta Army
Posts: 2597


WWW
Reply #7 on: September 27, 2007, 09:27:17 PM

Just in case some of you are using Excel 2007: Bug in Excel 2007

Basically some calculations that should come out to 65535 are showing as 100000.


I know what I should ask my salary to be.

Eternity is a very long time, especially towards the end.
Murgos
Terracotta Army
Posts: 7474


Reply #8 on: September 28, 2007, 07:40:10 AM

Code? Maybe, but what the hell would you use integers of 65k or higher for?  An array that big would break excel.

Finance calculations?  We work to four decimal places in unit prices - the likelihood of hitting 65535.00 is basically nil.

We usually use MiniTab but test data and process evaluation is handled pretty well with a spreadsheet.  Linear regressions (amongst many other tools) and such on a few million data points is pretty handy, especially when you can just use it and not have to have all kinds of project specific programs to support.

As I recall Excel 2007 was supposed to lift the limit on number of rows you could work with (MiniTab is limited to 10 million) though I don't think anyone here was seriously considering using it.

When you are doing calculations on millions of items there is a pretty good chance that one of them might end up being 65535.

"You have all recieved youre last warning. I am in the process of currently tracking all of youre ips and pinging your home adressess. you should not have commencemed a war with me" - Aaron Rayburn
Soln
Terracotta Army
Posts: 4737

the opportunity for evil is just delicious


Reply #9 on: September 28, 2007, 01:12:00 PM

buffer overflow... someone didn't increment their array properly    cheesy

in other news, "Better Uninstall: AOL Instant Messenger Has Giant Worm Hole".  Good times.   NDA
Valmorian
Terracotta Army
Posts: 1163


Reply #10 on: October 01, 2007, 07:19:09 AM

I thought about this and can't really see how this would break any of my spreadsheets. 

I'm not sure why 850 units times 77.10 would be such an impossible thing to encounter.
Lt.Dan
Terracotta Army
Posts: 758


Reply #11 on: October 01, 2007, 05:38:21 PM

In the spreadsheets I use the calc would have to be 850.0000 times 77.1000 since all the inputs are to four decimal places (no rounding) - that's one in a hundred million right there.  That's just the calc.  There is also error checking at an aggregate level so the total of all the units is added up and multiplied by the price of those units to get a check on the total of the individual transactions.

hal
Terracotta Army
Posts: 835

Damn kids, get off my lawn!


Reply #12 on: October 01, 2007, 06:53:26 PM

Um ..It is broke. Trust it only if you want to be fired.

I started with nothing, and I still have most of it

I'm not a complete idiot... Some parts are still on backorder.
Trippy
Administrator
Posts: 23657


Reply #13 on: October 10, 2007, 06:05:22 AM

Pages: [1] Go Up Print 
f13.net  |  f13.net General Forums  |  General Discussion  |  Topic: Math is hard - serious bug in Excel 2007  
Jump to:  

Powered by SMF 1.1.10 | SMF © 2006-2009, Simple Machines LLC