[R-sig-Geo] Help converting neighbor list to vector

Corey Sparks corey.sparks at utsa.edu
Wed Jun 23 14:04:19 CEST 2010

Hello all,
I have a neighbor list, formed from poly2nb in spdep, and a spatial  
weight matrix formed by nbmat(). I then extract all of the elements of  
the matrix that are nonzero (the neighbors for each observation),  
here's what i've done

sa.wt<-nb2mat(sa.nb, style="B")
x<-apply(sa.wt, 1,function(x){which(x==1)})

so, here is a preview of the x object from above:

[1]  4 11

[1]  3 10 12

[1]  2  7 10 12 13

If we look at the entire dataset (there are ~900 observations) the  
maximum number of neighbors is 10.  I would like to turn this 'x'  
object into a vector of form:

row1  obs#  obs#offirstneighbor  obs#ofsecondeighbor  
obs#ofthirdneighbor ... total#ofneighbors

like this example

      V1 V2  V3  V4 V5 V6
1     1 11   2   0  0  2
2     2 12   1   3  0  3
3     3 13   2   4  0  3

Here's what I've tried, but alas it fails

nbs<-matrix(0, nrow=dim(slot(sa,"data.frame"))[1], ncol=10)

for(i in 1:dim(slot(sa,"data.frame"))[1]) { for (j in length(x[[j]])) 

I've also tried unlist(x), but that doesn't give the result i'm  
looking for.  If anyone might have an idea on this one I'd really be  
appreciative, and i'd be happy to make the data available if anyone  
would like.

Corey Sparks
Assistant Professor
Department of Demography and Organization Studies
College of Public Policy
501 West Durango Blvd
Monterrey Building 2.270C
San Antonio, TX 78207
corey.sparks 'at' utsa.edu
210 458 3166

More information about the R-sig-Geo mailing list