"For" loop output issue in R

Question

I am struggling with an output allocation issue in a "for" loop: I generate an empty matrix ("results") with appropriate (is it not???) dimensions where the loop should allocate the output from the function (pls), but I am getting an error... Here is the code:

library(pls)
set.seed(10000)
mat <- replicate(100, rnorm(100))
y <- as.matrix(mat[,1], drop=F)
x <- mat[,2:100]
min.nc <- 3
max.nc <- 8
nc <- seq(min.nc, max.nc)
results <- matrix(NA, length(nc), 1)
for(q in nc) {
  pls <- plsr(y ~ x, ncomp = q)
  results[q,] <- RMSEP(pls, intercept=F, estimate="train")$val[1,1,q]
}

The error message after running the "for" loop says:

Error in `[<-`(`*tmp*`, q, , value = 0.329631604369802) : 
  subscript out of bounds

...which I don't understand, since the iterator "nc" has exactly the size of "length(nc)". Any ideas?

Many thanks for your time and help!

Chega

Justin · Accepted Answer · 2013-05-29 14:42:18Z

4

Your q is 3:8 and your matrix results is 1:6. Thus, when it tries to assign results[q,] when q > 6 you get this error.

> results[7,]
Error in results[7, ] : subscript out of bounds

The value of q is each value of nc rather than seq_along(nc).

for (q in nc) print(q)

Versus

for (q in seq_along(nc)) print(nc[q])

answered May 29, 2013 at 14:42

Justin

43.4k9 gold badges96 silver badges111 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Chega Over a year ago

So seq_along() is the tool I was obviously missing! Many thanks for your help!

Justin Over a year ago

The naive method would be 1:length(nc) but this will fail in surprising ways if nc is length 0 or doesn't exist.

Chega Over a year ago

I see - I can use seq_along() and seq_len() in most cases when 1:length() seems handy...

Collectives™ on Stack Overflow

"For" loop output issue in R

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related