R help - Jan 2007 - strange behaviour with equality after simple subtraction

hello,

 

today while trying to extract data from a list for subsequent analysis, i
stumbled upon this funny behavior on my system:

 
> x<-c(0.1,0.9)
> 1-x[2]
[1] 0.1
> x[1]
[1] 0.1
> x[1]==1-x[2]
[1] FALSE
> x[1]>1-x[2]
[1] TRUE

 
> x<-c(0.3,0.7)
> x[1]
[1] 0.3
> x[2]
[1] 0.7
> 1-x[2]
[1] 0.3
> x[1]==1-x[2]
[1] FALSE

 

but:

 
> x<-c(0.4,0.6)
> x[1]
[1] 0.4
> x[2]
[1] 0.6
> 1-x[2]
[1] 0.4
> x[1]==1-x[2]
[1] TRUE

 

any ideas why this might be? in my data i have values that are calculated
like above, e.g.

 

y=x if x<=0.5, else y=1-x 

 

now because of that behaviour whenever i want to get values e.g. <0.1 from
my data, i will also get all values of y=0.1 that came from the subtraction.


 

i'm using R 2.4.0 on windows XP; this is my sessionInfo():

 
> sessionInfo()
R version 2.4.0 (2006-10-03) 

i386-pc-mingw32 

 

locale:

LC_COLLATE=Spanish_Spain.1252;LC_CTYPE=Spanish_Spain.1252;LC_MONETARY=Spanis
h_Spain.1252;LC_NUMERIC=C;LC_TIME=Spanish_Spain.1252

 

attached base packages:

[1] "methods"   "stats"     "graphics" 
"grDevices" "utils"     "datasets"
"base"     
> 
 

thanks a lot

 

martin


	[[alternative HTML version deleted]]

"martin sikora" <martin.sikora at upf.edu> wrote:
> today while trying to extract data from a list for subsequent analysis, i
> stumbled upon this funny behavior on my system:
> 
> > x<-c(0.1,0.9)
> 
> > 1-x[2]
> 
> [1] 0.1
> 
> > x[1]
> 
> [1] 0.1
> 
> > x[1]==1-x[2]
> 
> [1] FALSE
> 
> > x[1]>1-x[2]
> 
> [1] TRUE
> 
Not at all strange, an expected property of floating-point
arithmetic and one of the most frequently asked questions here.
> print(0.1, digits=17)
[1] 0.1
> print(1 - 0.9, digits=17)
[1] 0.09999999999999998
> 
A simple description of the issue is at

http://docs.python.org/tut/node16.html

In most cases, it suffices to test for approximate difference or
relative difference. The former would look like this

if (abs(x[1] - x[2]) < eps)) ...
 
with "eps" set to something you think is an insignificant
difference, say 1.0e-10.


-- 
Mike Prager, NOAA, Beaufort, NC
* Opinions expressed are personal and not represented otherwise.
* Any use of tradenames does not constitute a NOAA endorsement.

FAQ on R 7.31. ?all.equal   ?identical  

Have you read these?

Bert Gunter
Genentech Nonclinical Statistics
South San Francisco, CA 94404


-----Original Message-----
From: r-help-bounces at stat.math.ethz.ch
[mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of Mike Prager
Sent: Friday, January 26, 2007 8:41 AM
To: r-help at stat.math.ethz.ch
Subject: Re: [R] strange behaviour with equality after simple subtraction

"martin sikora" <martin.sikora at upf.edu> wrote:
> today while trying to extract data from a list for subsequent analysis, i
> stumbled upon this funny behavior on my system:
> 
> > x<-c(0.1,0.9)
> 
> > 1-x[2]
> 
> [1] 0.1
> 
> > x[1]
> 
> [1] 0.1
> 
> > x[1]==1-x[2]
> 
> [1] FALSE
> 
> > x[1]>1-x[2]
> 
> [1] TRUE
> 
Not at all strange, an expected property of floating-point
arithmetic and one of the most frequently asked questions here.
> print(0.1, digits=17)
[1] 0.1
> print(1 - 0.9, digits=17)
[1] 0.09999999999999998
> 
A simple description of the issue is at

http://docs.python.org/tut/node16.html

In most cases, it suffices to test for approximate difference or
relative difference. The former would look like this

if (abs(x[1] - x[2]) < eps)) ...
 
with "eps" set to something you think is an insignificant
difference, say 1.0e-10.


-- 
Mike Prager, NOAA, Beaufort, NC
* Opinions expressed are personal and not represented otherwise.
* Any use of tradenames does not constitute a NOAA endorsement.

______________________________________________
R-help at stat.math.ethz.ch 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.