[R] Wrong computation of time differenze in POSIXct - additional digits

[peter dalgaard]

On Jun 15, 2012, at 20:33 , Rui Barradas wrote:

> Hello,
> 
> A classic of floating-point accuracy is
> 
> > 3/5 - 3/5
> [1] 0
> > 3/5 - (2/5 + 1/5)
> [1] -1.110223e-16
> > 3/5 - 2/5 - 1/5
> [1] -5.551115e-17
> 
> Rui Barradas
> 

Yes. There are only about 16 significant digits in (64 bit) floating point. One further point is that times are stored internally as seconds since Jan 1 1970, of which there has been quite a few by now:

> unclass(Sys.time())
[1] 1339793894

with already 10 digits before the decimal point, you can only expect fractional seconds to be accurate to about 6 digits.

-pd


> Em 15-06-2012 18:18, David Winsemius escreveu:
>> 
>> On Jun 15, 2012, at 12:49 PM, Julia wrote:
>> 
>>> Hello,
>>> 
>>> I wanted to compute the time differenze between to times:
>>> 
>>> first =as.POSIXct( "2012-06-15 16:32:39.0025 CEST")
>>> second  = as.POSIXct("2012-06-15 16:32:39.0086 CEST")
>>> second - first
>>> 
>>> The result is
>>> Time difference of 0.006099939 secs
>>> 
>>> instead of just 0.0061 secs
>>> So R adds aditional  numbers after the result.
>> 
>> It's a floating point representation issue.  You don't really want to
>> change that value, but are asking to see something different:
>> 
>> > round ( second - first, 4)
>> Time difference of 0.0061 secs
>> 
>>> I know I could round it in this case.
>>> But I am working with a large data set and need to always get the
>>> correct result.
>>> 
>>> difftime() does not work correct either.
>>> 
>>> Has anybody a suggestion how to get the correct result?
>> 
>> Use a computer system that runs on exact arithmetic?
>> 
>> Read FAQ 7.31
>> http://cran.r-project.org/doc/FAQ/R-FAQ.html#Why-doesn_0027t-R-think-these-numbers-are-equal_003f
>> 
>> 
>> (And expect to read about 4-6 similar messages in the next hour.)
>> 
>>> 
>>> Thank you
>>> Julia
>>> 
>>> ______________________________________________
>>> 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.
>> 
>> David Winsemius, MD
>> West Hartford, CT
>> 
>> ______________________________________________
>> 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.
> 
> ______________________________________________
> 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.

-- 
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com