R devel - May 2005 - rank (PR#7850)

Full_Name: Stella David
Version: 2.1.0
OS: mingw32
Submission from: (NULL) (137.250.161.136)


It seems that rank doesn't work right on vectors containing negative values.
If the negative value is the first component or there are only two components,
it works, but if the negative value is somewhere in the middle there's a
problem.

Here are my examples:
> # wrong:
> x<- c(2,-5,1,3);x
[1]  2 -5  1  3
> rank(x)
[1] 3 1 2 4
> x[rank(x)]
[1]  1  2 -5  3
> 
> # but this works:
> y<- c(-5,2,1,3);y
[1] -5  2  1  3
> rank(y)
[1] 1 3 2 4
> y[rank(y)]
[1] -5  1  2  3
> z <- c(3,-5);z
[1]  3 -5
> rank(z)
[1] 2 1
> z[rank(z)]
[1] -5  3
> 
> version
         _              
platform i386-pc-mingw32
arch     i386           
os       mingw32        
system   i386, mingw32  
status                  
major    2              
minor    1.0            
year     2005           
month    04             
day      18             
language R              

Stella David

You don't understand what rank() does: it is not the same as sort.list but 
rather the reverse (in the absence of ties).
>  x<- c(2,-5,1,3);x
[1]  2 -5  1  3
> y <- x
> y[rank(x)] <- x
> y
[1] -5  1  2  3

The ranks are the numbers in the ordering, and -5 is the smallest and so 
has rank 1, and so on.

Not a bug at all.

On Wed, 11 May 2005 stella.david@math.uni-augsburg.de wrote:
> Full_Name: Stella David
> Version: 2.1.0
> OS: mingw32
> Submission from: (NULL) (137.250.161.136)
>
>
> It seems that rank doesn't work right on vectors containing negative
values.
> If the negative value is the first component or there are only two
components,
> it works, but if the negative value is somewhere in the middle there's
a
> problem.
>
> Here are my examples:
>> # wrong:
>> x<- c(2,-5,1,3);x
> [1]  2 -5  1  3
>> rank(x)
> [1] 3 1 2 4
>> x[rank(x)]
> [1]  1  2 -5  3
>>
>> # but this works:
>> y<- c(-5,2,1,3);y
> [1] -5  2  1  3
>> rank(y)
> [1] 1 3 2 4
>> y[rank(y)]
> [1] -5  1  2  3
>> z <- c(3,-5);z
> [1]  3 -5
>> rank(z)
> [1] 2 1
>> z[rank(z)]
> [1] -5  3
>>
>> version
>         _
> platform i386-pc-mingw32
> arch     i386
> os       mingw32
> system   i386, mingw32
> status
> major    2
> minor    1.0
> year     2005
> month    04
> day      18
> language R
>
> Stella David
>
> ______________________________________________
> R-devel@stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
>
-- 
Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595