Tony, Sundar,
Thank you for your codes, they all work, but the "tricky" code of
Tony seems more efficient than the sapply() solution of Sundar,
especially for big size array. For a quick test:
> dm = matrix(rnorm(100000*40),ncol=40)
> idx=cbind(sample(1:ncol(dm), nrow(dm), replace=T),
sample(1:ncol(dm), nrow(dm),replace=T),
sample(1:ncol(dm), nrow(dm),replace=T))
> str(idx)
int [1:100000, 1:3] 6 2 1 6 24 23 21 36 31 21 ...
> system.time(a <- array(dm[cbind(as.vector(row(idx)), as.vector(idx))], dim=dim(idx)))
[1] 0.12 0.01 0.14 NA NA
> system.time(b <- t(sapply(seq(nrow(dm)), function(i) dm[i, idx[i, ]])))
[1] 1.60 0.00 1.69 NA NA
> all(a==b)
[1] TRUE
diana
______________________________________________
[EMAIL PROTECTED] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
