R help - May 2011 - constructing nxn matrices involving calculations& conditions based on other tables

Dear R-users,



I am having trouble constructing nxn matrices involving calculations&
conditions based on other tables.

Below I describe a simple example of what I am trying to do:



Given the Table A (n x m):





Species1

Species2

X

A1,1

A1,2

Y

A2,1

A2,2

Z

A3,1

A3,2



I want to create a matrix M (n xn)





X

Y

Z

X

-

M1,2

M1,3

Y

M2,1

-

M2,3

Z

M3,1

M3,2

-





Where M1,2 is equal to the number of common species between X & Y

So if:

 A1,1 = 5; A1,2 =0

A2, 1 = 1; A2,2 =3

Then M1,2 =  1



And also a matrix N (nxn)



A

B

C

A

-

N1,2

N1,3

B

N2,1

-

N2,3

C

N3,1

N3,2

-





Where N1,2 is  Sum A1,i – Sum A2,i



Any help or guidance on where I should look for information on these type of
calculations will be very appreciated!



Many thanks,



Luisa

	[[alternative HTML version deleted]]

Hi Luisa,

it was really difficult to manage to understand what you need. Assuming
that i've done it correctly this should be similar to what you want:

A.matrix <- matrix(rpois(n=28,lambda=2),nrow=7)

M.matrix <- matrix(0,nrow=dim(A.matrix)[2],ncol=dim(A.matrix)[2])
for(i in 1:dim(df)[2])
  for(j in i:dim(df)[2]) {
  prod <- df[,i]*df[,j]
  M.matrix[i,j] <- length(which(prod!=0))
  }

sums <- apply(A.matrix,2,sum)
N.matrix <-
matrix(0,nrow=dim(A.matrix)[2],ncol=dim(A.matrix)[2])
> for (i in 1:length(sums))
  for (j in i:length(sums)){
  diff <- sums[i]-sums[j]
  N.matrix[i,j] <- diff
  }

I briefly searched a way to to this without cycling inside objects, but
i wasn't able to do it straightforwardly. Probably somebody could
give a more efficient and simple hint.

Marco.