On 7/20/2006 3:49 PM, Steven McKinney wrote:
>
> Running R.app on Mac OS X 10.4
>
>> version
> _
> platform powerpc-apple-darwin8.6.0
> arch powerpc
> os darwin8.6.0
> system powerpc, darwin8.6.0
> status
> major 2
> minor 3.1
> year 2006
> month 06
> day 01
> svn rev 38247
> language R
> version.string Version 2.3.1 (2006-06-01)
>>
>
>
> I am trying to learn how to modify functions
> in a locked environment.
>
> For an example, suppose I'm using the package "zoo".
>
> zoo contains function "rollmean.default"
>
>> rollmean.default
> function (x, k, na.pad = FALSE, align = c("center", "left", "right"),
> ...)
> {
> x <- unclass(x)
> n <- length(x)
> y <- x[k:n] - x[c(1, 1:(n - k))]
> y[1] <- sum(x[1:k])
> rval <- cumsum(y)/k
> if (na.pad) {
> rval <- switch(match.arg(align), left = {
> c(rval, rep(NA, k - 1))
> }, center = {
> c(rep(NA, floor((k - 1)/2)), rval, rep(NA, ceiling((k -
> 1)/2)))
> }, right = {
> c(rep(NA, k - 1), rval)
> })
> }
> return(rval)
> }
> <environment: namespace:zoo>
>
> Suppose for whatever reason I want output to be
> in percent, so I'd like to modify the result to be
> rval <- 100 * cumsum(y)/k
>
> I cannot just copy the function and change it, as the namespace
> mechanism ensures the rollmean.default in 'zoo' continues to be used.
>
> If I use
> fixInNamespace("rollmean.default", ns = "zoo")
> I can edit the rval <- cumsum(y)/k line to read
> rval <- 100 * cumsum(y)/k
> save the file and exit the R.app GUI editor.
>
> But this does not update the exported copy of the
> function (the documentation for fixInNamespace says
> this is the case) - how do I accomplish this last step?
>
> If I list the function after editing, I see the original
> copy. But if I reinvoke the editor via fixInNamespace(),
> I do see my modification.
> Where is my copy residing? How do I push it out
> to replace the exported copy?
>
> Is this the proper way to modify a package function?
> Are there other ways? I've searched webpages, R news,
> help files and have been unable to find out how to
> get this process fully completed.
You could modify the zoo source and recompile it, or you could write
your own function that calls the one in zoo and modifies the result.
I don't think you should modify zoo without really careful thought: you
should assume that the package has been tested the way it was written,
and may give incorrect results if you go in and change one function
without considering everything else in the package.
So I'd recommend writing a wrapper, e.g.
myrollmean <- function(x, k, na.pad = FALSE, align = c("center", "left",
"right"), ...) {
result <- rollmean(x, k, na.pad, align, ...)
return(result*100)
}
> ...
Duncan Murdoch
>
> Any guidance appreciated.
>
>
> Steven McKinney
>
> Statistician
> Molecular Oncology and Breast Cancer Program
> British Columbia Cancer Research Centre
>
> email: [EMAIL PROTECTED]
>
> tel: 604-675-8000 x7561
>
> BCCRC
> Molecular Oncology
> 675 West 10th Ave, Floor 4
> Vancouver B.C.
> V5Z 1L3
> Canada
>
> ______________________________________________
> R-help@stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
______________________________________________
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
