R help - Nov 2012 - Ceiling function gives me wrong answer

Hi all,

I have a very simple code, but gives a wrong answer.

a=1000
b=1000^(1/3)
c=ceiling(a/b)

then c=101,

if change the code to be
a=1000
b=10
c=ceiling(a/b)

then c=100 is fine.

Thank you for the help.

	[[alternative HTML version deleted]]

> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at
r-project.org]
> On Behalf Of Ying Zheng
> Sent: Wednesday, November 28, 2012 6:11 PM
> To: r-help at r-project.org
> Subject: [R] Ceiling function gives me wrong answer
> 
> Hi all,
> 
> I have a very simple code, but gives a wrong answer.
> 
> a=1000
> b=1000^(1/3)
> c=ceiling(a/b)
> 
> then c=101,
> 
> if change the code to be
> a=1000
> b=10
> c=ceiling(a/b)
> 
> then c=100 is fine.
> 
> Thank you for the help.
> 
This is another variation on FAQ 7.31.  You are trying to calculate the cube 
root using a power of 1/3.  But 1/3 cannot be represented exactly in a finite
binary floating-point system.  If you print the value of b with enough digits
you will see that it is not equal to 10 because the representation of the value
1/3 ends up being slightly smaller that one-third.
> print(1/3, digits=20)
[1] 0.33333333333333331483
>
> b <- 1000^(1/3)
> print(b, digits=20)
[1] 9.9999999999999982236


Hope this is helpful,

Dan

Daniel Nordlund
Bothell, WA USA

not sure ceiling() is behaving inappropriately..
> b <- 1000^( 1/3 )
> b
[1] 10
> options( digits = 22 )
> b
[1] 9.999999999999998223643


check out
http://blog.revolutionanalytics.com/2009/11/floatingpoint-errors-explained.htmlfor
more detail







On Wed, Nov 28, 2012 at 9:10 PM, Ying Zheng
<martin.yingzheng@gmail.com>wrote:
> Hi all,
>
> I have a very simple code, but gives a wrong answer.
>
> a=1000
> b=1000^(1/3)
> c=ceiling(a/b)
>
> then c=101,
>
> if change the code to be
> a=1000
> b=10
> c=ceiling(a/b)
>
> then c=100 is fine.
>
> Thank you for the help.
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
	[[alternative HTML version deleted]]

On 29-11-2012, at 03:10, Ying Zheng wrote:
> Hi all,
> 
> I have a very simple code, but gives a wrong answer.
> 
> a=1000
> b=1000^(1/3)
> c=ceiling(a/b)
> 
> then c=101,
> 
> if change the code to be
> a=1000
> b=10
> c=ceiling(a/b)
> 
> then c=100 is fine.
See R FAQ 7.31 "Why doesn't R think these numbers are equal?"
> sprintf("%.17f",1000/10)
[1] "100.00000000000000000"
> sprintf("%.17f",1000/1000^(1/3))
[1] "100.00000000000001421"

Berend