R help - Jan 2007 - A vectorization question

Hi,
 
A function calculates the absolute difference between the two largest values
of each row of a matrix, as shown in the following example code:
 
cx <- matrix(runif(15),5)
cy <- t( apply(cx, 1, order, decreasing=TRUE) )
 
cz <- rep(0, nrow(cx))
for( i in 1:nrow(cx) ) cz[i] <- abs(diff(cx[i, cy[i,1:2]]))
 
Anybody has any ideas on how the last loop can be vectorized?

Thanks. 
 
Christos Hatzis, Ph.D.
Nuvera Biosciences, Inc.
400 West Cummings Park
Suite 5350
Woburn, MA 01801
Tel: 781-938-3830
www.nuverabio.com <http://www.nuverabio.com/>

On Tue, 2007-01-09 at 16:10 -0500, Christos Hatzis
wrote:
> Hi,
>  
> A function calculates the absolute difference between the two largest
values
> of each row of a matrix, as shown in the following example code:
>  
> cx <- matrix(runif(15),5)
> cy <- t( apply(cx, 1, order, decreasing=TRUE) )
>  
> cz <- rep(0, nrow(cx))
> for( i in 1:nrow(cx) ) cz[i] <- abs(diff(cx[i, cy[i,1:2]]))
>  
> Anybody has any ideas on how the last loop can be vectorized?
> 
> Thanks. 

How about this:
> mat <- matrix(sample(1:50, 12), ncol = 4)
> mat
     [,1] [,2] [,3] [,4]
[1,]   39    1   22   11
[2,]   34   28   13   48
[3,]   25   40   38    3

> apply(mat, 1, function(x) abs(diff(sort(x, decreasing = TRUE)[1:2])))
[1] 17 14  2

Or
> apply(mat, 1, function(x) diff(sort(x)[3:4]))
[1] 17 14  2

HTH,

Marc Schwartz