[R] Matrix Construction; Subdiagonal

[Bill.Venables at csiro.au]

This is horrible on memory if you are dealing with large matrices.

Here is a slightly more slick version of the function I have already posted:


xyz <- function (v, k) {
    n <- length(v) + abs(k)
    x <- matrix(0, n, n)
    i <- (1 - min(0, k)):(n - max(0,k))
    j <- (1 + max(0, k)):(n + min(0,k))
    x[cbind(i,j)] <- v
    x
}

The point of this is that it does not hold three copies of the matrix in memory at once (x, row(x) and col(x)).  It uses the matrix index idea instead.

If you were going to be doing this a lot, you would be better writing a function `Diag<-`, say, which places a vector at a specified diagonal position of a given matrix.  This is an easy exercise for the reader! 


Bill Venables
http://www.cmis.csiro.au/bill.venables/ 


-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf Of Sundar Dorai-Raj
Sent: Thursday, 12 March 2009 10:00 AM
To: Stu Field
Cc: r-help at r-project.org
Subject: Re: [R] Matrix Construction; Subdiagonal

You can always write your own function:

myDiag <- function(x, vec, k) {
  x[row(x) == col(x) - k] <- vec
  x
}

myDiag(A, vec, -1)

Of course, you should probably do some input checking too.

--sundar

On Wed, Mar 11, 2009 at 4:57 PM, Stu Field <sgf at colostate.edu> wrote:
> Sure, that'll work fine, thanks.
> But I guess I was looking for something more similar to MatLab, I'm really
> surprised R doesn't have a preset command for this (?)
> Thanks again,
> Stu
> On 11 . Mar . 2009, at 5:49 PM, Sundar Dorai-Raj wrote:
>
> Does this help?
>
> A <- matrix(0, 6, 6)
> vec <- 1:5
> A[row(A) == col(A) + 1] <- vec
>
> --sundar
>
> On Wed, Mar 11, 2009 at 4:42 PM, Stu Field <sgf at colostate.edu> wrote:
>
> I'm trying to enter a vector into the subdiagonal of a matrix but
>
> cannot find a command in R which corresponds to the MatLab version of
>
> diag(vec, k), where vec = the vector of interest, and k = the diagonal
>
> (k=0 for the diagonal; k=-1 for the subdiagonal; k=1 for
>
> superdiagonal, etc.)
>
> Is there an equivalent command in R?
>
> I'm looking for something like this:
>
> vec = seq(1, 5, 1)        # vector of interest
>
> A = xyz(vec,-1)           # creates a 6x6 matrix with vec on the
>
> subdiagonal
>
> where xyz is some function similar to diag, but with differing
>
> arguments.
>
> I can't believe there is not a simple way to do this...
>
> Thanks for your help,
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> Stu Field, PhD
>
> Postdoctoral Fellow
>
> Department of Biology
>
> Colorado State University
>
> 1878 Campus Delivery
>
> Fort Collins, CO 80523-1878
>
> Office: E208 Anatomy/Zoology
>
> Phone: (970) 491-5744
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
>
>
>
>
>        [[alternative HTML version deleted]]
>
> ______________________________________________
>
> R-help at 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.
>
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Stu Field, PhD
> Postdoctoral Fellow
> Department of Biology
> Colorado State University
> 1878 Campus Delivery
> Fort Collins, CO 80523-1878
> Office: E208 Anatomy/Zoology
> Phone: (970) 491-5744
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
>
>
>

______________________________________________
R-help at 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.