[R] S4 generic and method arguments -- "expanded signature"

EJ Nikelski nikelski at bic.mni.mcgill.ca
Sun Apr 25 22:46:18 CEST 2010

Hello R-List,

** This e-mail was initially bounced. Please forgive any duplicates.**

   I'm creating a new generic function and 3 associated methods, in
which each of the methods only needs a subset of the arguments
specified in the generic.

So, I create the generic like so (with the signature defaulting to all
of the args) ...

       def = function(filename,
                                       likeFile) {
standardGeneric("myFunction") }

In my first method, for example, I only need access to the first 4
args, so I try ...

dimSteps=NULL, dimStarts=NULL) { ... }

Now, when I build the package, I get the following warning ...

In method for function "myFunction": expanding the
signature to include omitted arguments in definition: likeTemplate =
"missing", likeFile = "missing"

I've read in "S4 Classes in 15 pages, more or less" (John Chambers),
that a method *can* contain a subset of the args in the generic, so
I'm uncertain why R is "expanding the signature".  I suspect that all
methods need to include all signature elements specified within the
generic, even if the signature elements are not in the method's arg
list ... is this correct?

At any rate, I do not want the signature to be expanded, since (1) I
don't want to add these 2 "missing" args to my Rd \alias's, and (2) I
don't want to have to add
these 2 missing args to the \S4method Rd entry, as it makes these
entries horribly long and does not really add any useful information
for the user.

 Perhaps I'm not setting up these methods optimally?  What I *want*
is to be able to have 3 methods that could be invoked as follows:

Method 1: myFunction(filename, dimLengths, dimSteps, dimStarts)
Method 2: myFunction(filename, likeTemplate="templete")   ->
Method 3: myFunction(filename, likeFile="file")   -> likeFile="character"

Note that the arguments for Methods 2 & 3 are both ("character", "character").
Does anyone have any suggestions that would both solve my S4 dispatch
problem, whilst keeping the Rd doc user friendly (i.e., not polluting
the entries with "missing" arguments)?

Thanks in advance for any suggestions.



> sessionInfo()
R version 2.10.1 (2009-12-14)

[1] en_CA.UTF-8/en_CA.UTF-8/C/C/en_CA.UTF-8/en_CA.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

Jim Nikelski, Ph.D.
Postdoctoral Research Fellow
Bloomfield Centre for Research in Aging
Lady Davis Institute for Medical Research
Sir Mortimer B. Davis - Jewish General Hospital
McGill University

More information about the R-help mailing list