[Rd] cannot use fanny in package cluster (PR#8830)

maechler at stat.math.ethz.ch maechler at stat.math.ethz.ch
Wed May 3 13:58:17 CEST 2006


>>>>> "UweL" == Uwe Ligges <ligges at statistik.uni-dortmund.de>
>>>>>     on Wed,  3 May 2006 08:39:33 +0200 (CEST) writes:

    UweL> ghuang at stat.nctu.edu.tw wrote:
    >> Full_Name: Guan-Hua Huang Version: 2.0.1 OS: Linux
    >> Submission from: (NULL) (140.113.114.123)
    >> 
    >> 
    >> I install the package cluster by using
    >> install.packages("cluster"). After install it, it runs
    >> fine for function clara, but it does not work for
    >> function fanny. I did the following things:
    >> 
    >> library(cluster) set.seed(21) x <- rbind(cbind(rnorm(10,
    >> 0, 0.5), rnorm(10, 0, 0.5)), cbind(rnorm(15, 5, 0.5),
    >> rnorm(15, 5, 0.5)), cbind(rnorm( 3,3.2,0.5), rnorm(
    >> 3,3.2,0.5))) .proctime00 <- proc.time() (fannyx <-
    >> fanny(x, 2))
    >> 
    >> and got the following messages:
    >> 
    >> Error in .Fortran("fanny", as.integer(n), as.integer(jp),
    >> k, x2, dis = dv, ; Fortran function name not in DLL for
    >> package cluster
    >> 
    >> Any help.
    >> 
    >> Guan-Hua

    UweL> If this is a bug, then with your installation of the
    UweL> package, but NOT with R nor with the package.

    UweL> Please read the FAQs what a bug is and how to report
    UweL> it (check with a ***recent*** version of R first!).

Indeed.

However, as Uwe properly pointed out in a related E-mail,
there is a bug -- still not in R , and hence not for R-bugs !! ---
in the DESCRIPTION file of newer versions of 'cluster':

You shouldn't have been able to install the current version of
the package 'cluster' with your ancient version of R, since
cluster effectively requires R version >= 2.2.1
(because earlier versions of R had a bug in their registration
 of Fortran symbols).

Hence, if you want to continue using your outdated version of R,
you should really use the old version cluster that was part of
your version of R...

Martin



More information about the R-devel mailing list