[BioC] hcluster crashes R environment

[Vincent Carey 525-2265]

> Hello,
>
> I am trying to apply the function "hcluster" (contained within the
> "amap" package) to data matrix with no variability between its columns.
> The example I am using is:
>
>
>   x <- data.frame(col1=c(20,30,45,7,6,5,4,3),col2=c(20,30,45,7,6,5,4,3),
> stringsAsFactors=FALSE)
>   hcluster(x, method="correlation", link = "complete")
>
>
> In this case, the error generated is fatal and the application (R
> environment) stops completely (crashes) without any control of the
> error. I do understand it in the way one can not apply the correlation
> to a vector with no variability, but in the following case:

I can confirm this on R 2.7.0 for macosx; you did not tell us the platform
you are using; please read the posting guide and supply sessionInfo()
for such queries.

amap is contributed by Antoine Lucas, as noted on the CRAN distribution
point, please notify him.  Clearly there is inadequate error management
through his call to .C. (Developers are encouraged to use .Call instead.)

>
>   x <- data.frame(col1=c(20,30,45,7,6,5,4,3),col2=c(20,30,45,7,6,5,7,7),
> stringsAsFactors=FALSE)
>   hcluster(x, method="correlation", link = "complete")
>
> There are rows without variability, and others (last two rows)
> containing some. When I use this example, the R environment does not
> crash. So, my question is: Why the "hcluster" funcions generates this
> fatal error when no variability is found in all the matrix rows ?
>
>
> Thanks in advance
>
> --
> Regards
> Henrique Proença
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor
>
-------------- next part --------------

The information transmitted in this electronic communica...{{dropped:18}}