[R] Error in `[.matrix.coo`(x, rw, cl) : Subscripts out of bound

[Vidhu Choudhary]

Hi All,
I am facing a strange problem.  I am making a R Package that internally run
various packages and summarizes the results from them.
I am getting this error from the piece of code written below. The text in
red color is throwing
Error in `[.matrix.coo`(x, rw, cl) : Subscripts out of bound
Value of m is 99
This code works fine when I tested it. But money I make  library it gives me
error.
I have attached the file names messages.txt which contains the
warnings/messages generated while making R package.
I am out of ideas
Kindly suggest something

regards
vidhu
############################################################################
`sm.quantreg` <-
function(y, lambda = 1.0, tau = 0.5, interpolate = FALSE)  {
  nona <- !is.na(y)

  ## Sparse matrices: E, D, and B
  if(!interpolate)  {
    y.nona <- y[nona]
    m <- length(y.nona)  *1.0
    E <- as.matrix.csr(0,m,m)
    diag(E) <- 1
*    D <- diff(E)*
    B <- rbind(E, lambda * D)
   ystar <- c(y.nona, rep(0,m-1))

  }else  {
    ## interpolate missing values via weighting
    y[!nona] <- 0
    m <- length(y) *1.0
    E <- as.matrix.csr(0,m,m)

    diag(E) <- 1
    D <- diff(E)
    tmp <- as.matrix.csr(0,m,m)
    diag(tmp) <- as.numeric(nona)
    B <- rbind(tmp, lambda * D)
    ystar <- c(y, rep(0,m-1))
  }

  ## use sparse matrices and accompanying mathematical operations
  cat ("before try\n")
  try(myrq <- rq.fit.sfn(as.matrix.csr(B), ystar, tau))
  cat ("after try\n")
  if(exists("myrq"))  {
    if(!interpolate)  {
      ys <- rep(NA, length(y))
      ys[nona] <- myrq$coef
    }else  {
      ys <- myrq$coef
    }
  }else  {
    cat("Error triggered within Quantreg algorithm!\n")
    ys <- NULL
  }
  return(ys)
}

##########################
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: messages.txt
URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20080520/0e63a0f9/attachment.txt>