Linux:
>
x<-c("0","1");y<-c("a","-1");
x<y
[1] TRUE TRUE
Mac:
x<-c("0","1");y<-c("a","-1");
x<y
[1] TRUE FALSE
On both systems, LC_COLLATE/LC_CTYPE/LANG are set to en_US.UTF-8
In Java,FWIW, I get the Mac answer if I use String.compareTo and the
Linux answer if I use Collator.compareTo, but the result is consistent
on Mac and Linux.
Mick Jordan
On 3/11/16 6:52 PM, Mick Jordan wrote:> Linux: > > x<-c("0","1");y<-c("a","-1"); x<y > [1] TRUE TRUE > > Mac: > > x<-c("0","1");y<-c("a","-1"); x<y > [1] TRUE FALSE > > > On both systems, LC_COLLATE/LC_CTYPE/LANG are set to en_US.UTF-8 > > In Java,FWIW, I get the Mac answer if I use String.compareTo and the > Linux answer if I use Collator.compareTo, but the result is consistent > on Mac and Linux. >So this is probably related to the fact that the Linux system in question answers "ICU not in use" whereas the Mac answers "root"? Mick
> On Mar 11, 2016, at 6:52 PM, Mick Jordan <mick.jordan at oracle.com> wrote: > > Linux: > > x<-c("0","1");y<-c("a","-1"); x<y > [1] TRUE TRUE > > Mac: > > x<-c("0","1");y<-c("a","-1"); x<y > [1] TRUE FALSE >On a Mac (and noting the the documentation [somewhere that I cannot find right now] warns us the the collation sequence for characters is OS dependent): require(R.oo) charToInt("-") #[1] 45 charToInt("1") #[1] 49> > On both systems, LC_COLLATE/LC_CTYPE/LANG are set to en_US.UTF-8 > > In Java,FWIW, I get the Mac answer if I use String.compareTo and the Linux answer if I use Collator.compareTo, but the result is consistent on Mac and Linux. > > Mick Jordan > > ______________________________________________ > R-devel at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-develDavid Winsemius Alameda, CA, USA