Martin, that helps. I
I made the mistake you outlined below
> setGeneric("as.data.frame",
function(x) standardGeneric("as.data.frame"))
Creating a generic for 'as.data.frame' in package '.GlobalEnv'
(the supplied definition differs from and overrides the implicit generic
in package 'base': Formal arguments differ: (x), (x, row.names,
optional, ...))
[1] "as.data.frame"
then I'm in trouble -- I've created a generic 'as.data.frame', but since
the signature of my generic differs from the signature of
base::as.data.frame, the default behavior does NOT create a
as.data.frame,ANY method.
################
this, as you note leads to trouble
And of course got in trouble.
setClass(Class="ftpdir",
representation = list(dirframe="data.frame"))
if (!isGeneric("as.data.frame")) {
setGeneric("as.data.frame", function(dirlist )
standardGeneric("as.data.frame"))
}
setMethod("as.data.frame",
signature(dirlist = "ftpdir" ),
function (dirlist )
{
return( dirlist@dirframe)
}
)
On Thu, Jan 13, 2011 at 10:33 AM, Martin Morgan <[email protected]> wrote:
> On 01/13/2011 09:49 AM, steven mosher wrote:
> >
> >
> > On Thu, Jan 13, 2011 at 6:46 AM, Martin Morgan <[email protected]
> > <mailto:[email protected]>> wrote:
> >
> > On 01/12/2011 10:54 PM, steven mosher wrote:
> > > I have J Chambers wonderful text ( Software for data Analysis) and
> > I've been
> > > trying
> > > my hand at some very routine S4 OOP development.
> > >
> > > One of the things I was trying to do was to create some very basic
> S4
> > > classes. The first
> > > was simply a class that had a data.frame as part of its
> > representation.
> > >
> > > setClass("df",representation(dirframe="data.frame"))
> > >
> > > The object basically contains a data.frame that represents a file
> > directory
> > > listing
> > > with a column named filename, size, time, etc.
> > >
> > > And then I have methods for doing various things with this object.
> > >
> > > I then tried to tackle the problem of coercing this S4 object to a
> > > data.frame. Again just a learning exercise.
> > >
> > > The goal would be able to make a call like this
> > >
> > > testFrame <- as.data.frame(x)
> > >
> > > where x, was an object of class "df"
> > >
> > > If I try to define "as.data.frame" as a S4 method, then I can make
> > it work,
> > > but I then destroy the S3 functionality
> > > of as.data.frame, so that if I were to try to coerce a matrix to a
> > > data.frame it would work.
> >
> > Hi Steven --
> >
> > This works for me
> >
> > setClass("A", representation=representation(df="data.frame"))
> >
> > setMethod("as.data.frame", "A",
> > function(x, row.names=NULL, optional=FALSE, ...)
> > {
> > ## implementation, e.g.,
> > callGeneric(x@df, row.names=row.names, optional=optional, ...)
> > })
> >
> >
> > this makes no sense to me.
> >
> > Looking at this in the manual:
> > "A call to callGeneric can only appear inside a method definition. It
> > then results in a call to the current generic function. The value of
> > that call is the value of callGeneric. While it can be called from any
> > method, it is useful and typically used in methods for group generic
> > functions."
>
> callGeneric is not necessary, and I should just have said
>
> as.data.frame(x@df, row.names=row.names, optional=optional, ...)
>
> > I'm further confused. what is the "current" generic function?
>
> I think of a generic 'foo' as a function, and inside that function are
> functions foo,A-method, foo,B-method etc for classes A, B, .... When in
> one of these methods, foo,A-method, the current generic is the function
> 'foo'. For as.data.frame, the generic is unambiguous ('as.data.frame' !)
> and the interesting case are the group generics (?Logic, for instance),
> where a single method
>
> setMethod("Logic", function(e1, e2) callGeneric(<YOUR CODE HERE>))
>
> and you'll have in effect written 'methods' for all the operators
> defined in the 'Logic' group. Cool.
>
> >
> > > as.data.frame(new("A"))
> > Object of class "data.frame"
> > data frame with 0 columns and 0 rows
> > > as.data.frame(matrix(0, 3, 5))
> > V1 V2 V3 V4 V5
> > 1 0 0 0 0 0
> > 2 0 0 0 0 0
> > 3 0 0 0 0 0
> >
> >
> > Maybe you call setGeneric (no need to, setMethod will promote
> > as.data.frame automatically) in a way that does not specify the
> default
> > (arg useAsDefault) correctly?
> >
> >
> > I think that may have been the mistake.. what do you mean by
> > no need to call setGeneric?
>
> The only code I had was what was written above -- setClass and
> setMethod. I could also have
>
> setGeneric("as.data.frame")
>
> and would have been ok -- the default behavior of setGeneric in this
> case is to make a generic function as.data.frame, AND a method
> as.data.frame,ANY-method. The as.data.frame,ANY-method is implemented as
> base::as.data.frame, and is where objects not handled by methods I
> implement might end up being dispatched to. In a new R session, try
>
> setGeneric("as.data.frame")
> showMethods(as.data.frame)
> selectMethod(as.data.frame, "ANY")
>
> If I had done
>
> > setGeneric("as.data.frame",
> function(x) standardGeneric("as.data.frame"))
> Creating a generic for 'as.data.frame' in package '.GlobalEnv'
> (the supplied definition differs from and overrides the implicit generic
> in package 'base': Formal arguments differ: (x), (x, row.names,
> optional, ...))
> [1] "as.data.frame"
>
> then I'm in trouble -- I've created a generic 'as.data.frame', but since
> the signature of my generic differs from the signature of
> base::as.data.frame, the default behavior does NOT create a
> as.data.frame,ANY method.
>
> Not sure if this helps or not...
>
> Martin
>
> >
> > I really like chambers book, but there are certain parts where the lack
> > of simple examples
> > really makes it difficult to follow.
> >
> >
> > Martin
> >
> >
> > >
> > >
> > > So, I guess my question is what do I do, write an s3 method for
> > > as.data.frame that takes a "df" object as a paramter?
> > > The book wasn't exactly clear ( or I'm not that bright), or is
> > there a way
> > > to make the S4 method I wrote "as.data.frame"
> > > call the S3 method if needed?
> > >
> > > [[alternative HTML version deleted]]
> > >
> > > ______________________________________________
> > > [email protected] <mailto:[email protected]> 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.
> >
> >
> > --
> > Computational Biology
> > Fred Hutchinson Cancer Research Center
> > 1100 Fairview Ave. N. PO Box 19024 Seattle, WA 98109
> >
> > Location: M1-B861
> > Telephone: 206 667-2793
> >
> >
>
>
> --
> Computational Biology
> Fred Hutchinson Cancer Research Center
> 1100 Fairview Ave. N. PO Box 19024 Seattle, WA 98109
>
> Location: M1-B861
> Telephone: 206 667-2793
>
[[alternative HTML version deleted]]
______________________________________________
[email protected] 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.
