# [R] extract a subset of non-contiguous elements of a matrix

Bart Kastermans kasterma at kasterma.net
Thu Jun 19 15:24:45 CEST 2014

```On 19 Jun 2014, at 15:16, carol white <wht_crl at yahoo.com> wrote:

>  tm.1=rbind(c(1,-3,2,-4), c(1,-3,2,-4),c(1,-3,2,-4))
>
> > which(tm.1 > 0, arr.ind=TRUE)
>      row col
> [1,]   1   1
> [2,]   2   1
> [3,]   3   1
> [4,]   1   3
> [5,]   2   3
> [6,]   3   3
>
> so the answer should have the elements of tm.1 with the following indexes
> 1,1 1,3
> 2,1 2,3
> 3,1 3,3

In that example:

> t(apply(tm.1, 1, function(x) x[x>0]))
[,1] [,2]
[1,]    1    2
[2,]    1    2
[3,]    1    2

> class(t(apply(tm.1, 1, function(x) x[x>0])))
[1] "matrix"

>
>
> On Thursday, June 19, 2014 3:08 PM, Bart Kastermans <kasterma at kasterma.net> wrote:
>
>
> If you give an example of input and desired output I can think about this.  But at this
> point I do not understand what you want.  In the example I gave the positive elements do
> not form a submatrix in any way I can think of.
>
> On 19 Jun 2014, at 15:04, carol white <wht_crl at yahoo.com> wrote:
>
> > well it gives a vector which is useless as I want a matrix.
> >
> >
> > On Thursday, June 19, 2014 2:40 PM, Bart Kastermans <kasterma at kasterma.net> wrote:
> >
> >
> > > tm.1 <- matrix(c(11,22,33,-4), ncol=2)
> > > which(tm.1 > 0, arr.ind=TRUE)
> >  row col
> > [1,]  1  1
> > [2,]  2  1
> > [3,]  1  2
> > > tm.1[which(tm.1 > 0, arr.ind=TRUE)]
> > [1] 11 22 33
> >
> > This last command does what you ask I think.
> >
> > On 19 Jun 2014, at 14:12, carol white <wht_crl at yahoo.com> wrote:
> >
> > > The extracted values don't form a matrix and that's the question how to extract because which returns the indexes? that is, from
> > > 1,1
> > > 2,1
> > > 1,2
> > >
> > > how to retrieve values?
> > >
> > > Or if at the position 2,1, there is a negative value, how to retrieve
> > > 1,1
> > > 1,2
> > >
> > >
> > > Carol
> > >
> > >
> > > On Thursday, June 19, 2014 1:29 PM, Bart Kastermans <kasterma at kasterma.net> wrote:
> > >
> > >
> > > On 19 Jun 2014, at 13:19, carol white <wht_crl at yahoo.com> wrote:
> > >
> > > > Hi,
> > > > Is there a way to extract a subset of non-contiguous elements of a matrix elegantly and with 1 or very few scripts?
> > > >
> > > > Suppose I have a matrix of positive and negative numbers (m) and I want to retrieve only the positive number. This I can do
> > > >
> > > > which(m>0, arr.ind=T) which gives the indices of positive elements like (37,1), (80,1), ..., (54,2) etc. How can I extract positive numbers without looping on the indexes provided by which to make a new matrix?
> > >
> > > What matrix do you want?  For e.g.
> > >
> > > > tm.1 <- matrix(c(11,22,33,-4), ncol=2)
> > > > which(tm.1 > 0, arr.ind=TRUE)
> > >    row col
> > > [1,]  1  1
> > > [2,]  2  1
> > > [3,]  1  2
> > > > tm.1[which(tm.1 > 0, arr.ind=TRUE)]
> > > [1] 11 22 33
> > >
> > > The extracted values do not form a matrix.
> > >
> > > Either the above contains the answer, or I don’t understand the question.
> > >
> > > Best,
> > > Bart
> > >
> > >
> > > >
> > > > thanks,
> > > >
> > > > Carol
> > >
> > >
> >
> >
>
>

```