R help - Jun 2012 - Inconsistency using seq

Hi all,

Is there any problem of precision when using seq?. For example:

x<- seq(0,4,0.1)
x[4]=0.3

BUT:

x[4]-0.3=5.551115e-17

It means when I use this condition within an if clause, it does not find
values with 0.3 for x[4] as it is not precisely 0.3.
Is there any bug in seq() ?


--
View this message in context:
http://r.789695.n4.nabble.com/Inconsistency-using-seq-tp4633739.html
Sent from the R help mailing list archive at Nabble.com.

Le lundi 18 juin 2012 ? 12:58 -0700, hamoreno a ?crit :
> Hi all,
> 
> Is there any problem of precision when using seq?. For example:
> 
> x<- seq(0,4,0.1)
> x[4]=0.3
> 
> BUT:
> 
> x[4]-0.3=5.551115e-17
> 
> It means when I use this condition within an if clause, it does not find
> values with 0.3 for x[4] as it is not precisely 0.3.
> Is there any bug in seq() ?
No. Have a look at the FAQ:
http://cran.r-project.org/doc/FAQ/R-FAQ.html#Why-doesn_0027t-R-think-these-numbers-are-equal_003f


Cheers

Please read R FAQ 7.31. The problem is with your computer (and
everyone else's), and has nothing to do with seq().

On Mon, Jun 18, 2012 at 3:58 PM, hamoreno <hamoreno at asu.edu>
wrote:
> Hi all,
>
> Is there any problem of precision when using seq?. For example:
>
> x<- seq(0,4,0.1)
> x[4]=0.3
>
> BUT:
>
> x[4]-0.3=5.551115e-17
>
> It means when I use this condition within an if clause, it does not find
> values with 0.3 for x[4] as it is not precisely 0.3.
> Is there any bug in seq() ?
>
-- 
Sarah Goslee
http://www.functionaldiversity.org

No, this is rather the nature of floating point calculations. 

You may perhaps be looking for ?all.equal or R FAQ 7.31 (I think that's the
one) which is google-able. It's a complicated subject, but those should get
you started.

Best,
Michael

On Jun 18, 2012, at 2:58 PM, hamoreno <hamoreno at asu.edu> wrote:
> Hi all,
> 
> Is there any problem of precision when using seq?. For example:
> 
> x<- seq(0,4,0.1)
> x[4]=0.3
> 
> BUT:
> 
> x[4]-0.3=5.551115e-17
> 
> It means when I use this condition within an if clause, it does not find
> values with 0.3 for x[4] as it is not precisely 0.3.
> Is there any bug in seq() ?
> 
> 
> --
> View this message in context:
http://r.789695.n4.nabble.com/Inconsistency-using-seq-tp4633739.html
> Sent from the R help mailing list archive at Nabble.com.
> 
> ______________________________________________
> R-help at 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.

Thanks everyone... Seems that I will have to use round before seq to make
sure everything has the correct precision.

--
View this message in context:
http://r.789695.n4.nabble.com/Inconsistency-using-seq-tp4633739p4633750.html
Sent from the R help mailing list archive at Nabble.com.

actually, you should build your sequences with integers and scale those to get
floating point sequences.
---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live
Go...
                                      Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
--------------------------------------------------------------------------- 
Sent from my phone. Please excuse my brevity.

hamoreno <hamoreno at asu.edu> wrote:
>Thanks everyone... Seems that I will have to use round before seq to
>make
>sure everything has the correct precision.
>
>--
>View this message in context:
>http://r.789695.n4.nabble.com/Inconsistency-using-seq-tp4633739p4633750.html
>Sent from the R help mailing list archive at Nabble.com.
>
>______________________________________________
>R-help at 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.

On Mon, Jun 18, 2012 at 12:58:13PM -0700, hamoreno
wrote:
> Hi all,
> 
> Is there any problem of precision when using seq?. For example:
> 
> x<- seq(0,4,0.1)
> x[4]=0.3
> 
> BUT:
> 
> x[4]-0.3=5.551115e-17
> 
> It means when I use this condition within an if clause, it does not find
> values with 0.3 for x[4] as it is not precisely 0.3.
Hi.

Using round() is reasonably secure.

  x <- round(seq(0,4,0.1), digits=7)
  x[4] == 0.3
  [1] TRUE

None of the compared nunbers is 0.3, but they are both rounded
to the closest representable number to 0.3. 

  print(x[4], digits=20)

  [1] 0.2999999999999999889

  print(0.3, digits=20)

  [1] 0.2999999999999999889

See also

  http://rwiki.sciviews.org/doku.php?id=misc:r_accuracy

for some more hints related to this.

Petr Savicky.