[R] resizing data

[Brian Diggs]

On 1/25/2013 2:29 PM, emorway wrote:
> I played around with your example on the smaller dataset, and it seemed like
> it was doing what I wanted.  However, applying it to the larger problem, I
> didn't  get a resized 2D dataset that preserved the order I was hoping for.
> Hopefully the following illustrates the larger problem:
>
> x<-matrix(0,nrow=29328,ncol=7)
>
> # Now set the 70,000th element to 1 to find out where it ends up?
> # Iterating on columns first, then rows, the 70,000th element is
> # at index [10000,7]
>
> x[10000,7]<-1
> y<-t(matrix(x,nrow=546))
> dim(y)
> # [1] 376 546
>
> # Does 1 appear at index [129,112] as I expect
> # (128 complete rows x 546 cols = 69888 + 112 = 70,000) thus, row 129, col
> 112
> y[129,112]
> # [1] 0
>
> # No, so where is it?
> grep(1,y)
> # [1] 123293
>
> Where is that?

which(y==1, arr.ind=TRUE)
#      row col
# [1,] 341 328

You need an extra transposition

y <- t(matrix(t(x), nrow=546))
y[129,112]
# [1] 1
which(y==1, arr.ind=TRUE)
#      row col
# [1,] 129 112

-- 
Brian S. Diggs, PhD
Senior Research Associate, Department of Surgery
Oregon Health & Science University