R devel - Mar 2004 - "==" gives incorrect answers! (PR#6691)

Full_Name: Dong-Yun Kim
Version: 1.7.0/1.8.1
OS: WinMe/WinXP
Submission from: (NULL) (4.33.147.192)


Hi all,

It appears that R has a problem with evaluating logical "==" operator.
The
following are a few examples I found:
> 2.1+0.2==2.3
[1] FALSE
> 2.7+0.1==2.8
[1] FALSE
> 2.7+0.2==2.9
[1] FALSE
> 3.1+0.2==3.3
[1] FALSE

Thanks.

Dong-Yun Kim

dongyunhkim@yahoo.com wrote:
> Full_Name: Dong-Yun Kim
> Version: 1.7.0/1.8.1
> OS: WinMe/WinXP
> Submission from: (NULL) (4.33.147.192)
> 
> 
> Hi all,
> 
> It appears that R has a problem with evaluating logical "=="
operator. The
> following are a few examples I found:
> 
> 
>>2.1+0.2==2.3
> 
> [1] FALSE
> 
>>2.7+0.1==2.8
> 
> [1] FALSE
> 
>>2.7+0.2==2.9
> 
> [1] FALSE
> 
>>3.1+0.2==3.3
> 
> [1] FALSE
> 
> Thanks.
> 
> Dong-Yun Kim
> 
> ______________________________________________
> R-devel@stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
Not a bug!

These are (expected!) numerical issues of the representation of those 
values. Please read help("==") which tells you:

"For numerical values, remember == and != do not allow for the finite 
representation of fractions, nor for rounding error. Using all.equal 
with identical is almost always preferable."

And ?all.equal and ?identical tell you to use:

  identical(all.equal(2.7 + 0.1, 2.8), TRUE)


Uwe Ligges

dongyunhkim@yahoo.com writes:
> Full_Name: Dong-Yun Kim
> Version: 1.7.0/1.8.1
> OS: WinMe/WinXP
> Submission from: (NULL) (4.33.147.192)
> 
> 
> Hi all,
> 
> It appears that R has a problem with evaluating logical "=="
operator. The
> following are a few examples I found:
> 
> > 2.1+0.2==2.3
> [1] FALSE
> > 2.7+0.1==2.8
> [1] FALSE
> > 2.7+0.2==2.9
> [1] FALSE
> > 3.1+0.2==3.3
> [1] FALSE
> 
> Thanks.
> 
> Dong-Yun Kim
Please, do read up on elementary computer theory, in particular the
parts about inexact representation of decimal numbers as floating
point numbers, and why equality of such numbers cannot be trusted!

What you're seeing is the moral equivalent of 

1/3 + 1/6 = 0.333 + 0.666 = 0.999 != 1

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk)             FAX: (+45) 35327907