.C has a named argument 'PACKAGE', which you didn't set. I usually have such an error when I forget about the argument.
> -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Toralf Kirsten > Sent: Tuesday, June 22, 2004 9:53 AM > To: [EMAIL PROTECTED] > Subject: [Rd] function not in load table > > Hi, > I apologize for this often/old question. I found some hints > but couldn't solve the problem so far. > > I have C functions (incl. the header files) as well as the R > wrapper functions which I want to use for faster > calculations. These functions are included in a R package. > The installation process seems to be ok (no errors). I also > can load the package without errors. But when I call the > function I got the following error > wy.result <- > wy.grps(data1=X1, grpdata=groups, nres=10000, alpha1=0.05, > alpha2=0.05) Error in .C("wy_grps_R", as.double(X), > as.integer(n1), as.integer(n2), : > C function name not in load table Execution halted > > The parameter are > data1 - result of read.table() > grpdata - dito > nres - integer > alpha1 nad alpha1 - factors (float) > > In the R function wy.grps(...) the C function is called by > using the statement > > result <- .C("wy_grps_R", > as.double(X), > as.integer(n1), > as.integer(n2), > as.integer(p), > as.integer(unlist(grpidx)), > as.integer(grplen), > as.integer(grpnum), > as.character(WYFUN), > as.double(alpha2), > as.character(MINMAXFUN), > WYdist=double(nres), > as.integer(nres), > test.value=double(grpnum), > p.value=double(grpnum)) > > > My .First.lib.R is as follows: > .First.lib <- function(libname, pkgname) { > library.dynam("izbi", package = pkgname, lib.loc = libname) > data(COLS, package=pkgname) > data(ROWS, package=pkgname) > if (exists(".Dyn.libs")) remove(".Dyn.libs") > if (interactive() && getenv("DISPLAY") != "") x11() } > > I read something about R_CMethodDef in "Writing R Extensions" > but I'm not really sure where I should write it, may in the > .First.lib.R or in a separate file? > Any other possible mistakes? > > We are using R 1.9.0 on Fedora Linux. > > Thanks a lot, Toralf > > ______________________________________________ > [EMAIL PROTECTED] mailing list > https://www.stat.math.ethz.ch/mailman/listinfo/r-devel > > ______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
