[Rd] Re: R2.0.0 bug in function vcov in library survival (PR#7266)

Prof Brian Ripley ripley at stats.ox.ac.uk
Wed Oct 6 20:11:53 CEST 2004


On Wed, 6 Oct 2004, Thomas Lumley wrote:

> On Wed, 6 Oct 2004, Prof Brian Ripley wrote:
> 
> > Those vcov methods were removed from stats at
> >
> > 2004-08-28 20:40:13 +0100 survival now has vcov.{coxph,survreg}
> >
> > Looks like they have been removed from survival since: that is where they
> > should be.  They are very simple --
> >
> > vcov.coxph <- vcov.survreg <- function (object, ...) object$var
> >
> 
> No, they haven't been updated to cope with lazy loading (they are defined 
> in a top-level if statement to prevent conflicts in older versions of R).
> 
> vcov.coxph is actually
>      vcov.coxph<-function (object, ...) {
>          rval<-object$var
>          dimnames(rval)<-list(names(coef(object)),names(coef(object)))
>          rval
>      }
> since object$var doesn't have dimnames.

As Torsten has suggested, it looks like they were neither registered nor
exported from the survival namespace, but the objects are in the
namespace.  (Nothing to do with lazy-loading, then.)

> library(survival)
Loading required package: splines
> ls(asNamespace("survival"), patt="^vcov")
[1] "vcov.coxph"   "vcov.survreg"

Please register them in the next update.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595



More information about the R-devel mailing list