Gents:
The "eval(parse(...))" construction should almost always be avoided: it is
basically a misuse of R. There are exceptions, I suppose, but this does not
appear to be one of them.
Note that the use of numeric indexing does appear to be slightly faster
than logical indexing here, although I would say not enough to make a
practical difference. In any case:
A <- mat1New ## saves me a bit of editing
>
> system.time({
+ j40<- n*seq_len(nrow(A)/n)
+ vec1<- rep("j40",n)
+ res<- eval(parse(text=
paste(paste0("A","[",paste0(vec1,"-",seq(n)-1),",]"),collapse="*")
+ ))
+ })
user system elapsed
0.02 0.00 0.01
> system.time({
+ j <- seq_len(nrow(A))%%n
+ b <- A[j==0,]
+ for(i in seq_len(n-1))b <- b*A[j==i,]
+ })
user system elapsed
0.25 0.00 0.27
> system.time({
+ j <- n*seq_len(nrow(A)/n)
+ b1 <- A[j,]
+ for(i in seq_len(n-1))b1 <- b1*A[j-i,]
+ })
user system elapsed
0.01 0.00 0.02
## One should not invest too much faith in such superficial timing tests,
however.
I will have no further comments.
Cheers,
Bert
[[alternative HTML version deleted]]
______________________________________________
R-help@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.
