[Rd] Looks like a bug in subsetting of a complicated object

Kjetil Halvorsen kjetilbrinchmannhalvorsen at gmail.com
Thu Sep 2 00:22:12 CEST 2010


I don't understand what is happening! I have a (large) object sim1, an
matrix list
with dim c(101,101) where each element is an 3*3 matrix. I am
subsetting that with
a matrix coo, of dim c(100,2), of unique indices, but the resulting object
has length 99, not 100 as expected.

Code reproducing the problem follows:


library(RandomFields)

set.seed(123)
sim0 <- GaussRF(x=seq(0, 100, by=1),
                y=seq(0, 100, by=1),
                grid=TRUE, model="spherical",
                param=c(0, 1, 0, 10), trend=NULL, n=9, gridtriple=FALSE)

simmatrices <- function(arr) # arr must be an array of rank 3
                        {
               d <- dim(arr)
               n1 <- d[1]; n2 <- d[2] # we suppose d[3]==9
               res <- vector(length=n1*n2, mode="list")
               dim(res) <- c(n1, n2)
               for (i in 1:n1) for (j in 1:n2) {
                 x1 <- arr[i, j, 1:3];x2 <- arr[i, j, 4:6]
                 x3 <- arr[i, j, 7:9]
                 res[[i, j]] <- x1%o%x1 + x2%o%x2 + x3%o%x3
               }
               res
             }

sim1 <- simmatrices(sim0)

set.seed(234)
x <- sample(seq(0, 100, by=1), 101, replace=TRUE)
y <- sample(seq(0, 100, by=1), 101, replace=TRUE)
coo <- cbind(x=x, y=y)
coo <- unique(coo)

sim1.obs <- sim1[coo]

dim(coo)
length(sim1.obs)


> sessionInfo()
R version 2.11.1 (2010-05-31)
i686-pc-linux-gnu

locale:
 [1] LC_CTYPE=en_US.utf8       LC_NUMERIC=C
 [3] LC_TIME=en_US.utf8        LC_COLLATE=en_US.utf8
 [5] LC_MONETARY=C             LC_MESSAGES=en_US.utf8
 [7] LC_PAPER=en_US.utf8       LC_NAME=C
 [9] LC_ADDRESS=C              LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.utf8 LC_IDENTIFICATION=C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

other attached packages:
[1] lattice_0.18-8      gstat_0.9-69        sp_0.9-66
[4] RandomFields_1.3.41

loaded via a namespace (and not attached):
[1] fortunes_1.3-7 grid_2.11.1    tools_2.11.1
>

Kjetil



More information about the R-devel mailing list