[R] classes and methods

[mtmorgan at fhcrc.org]

Hi Edna --

Quoting Edna Bell <edna.bell01 at gmail.com>:

> I have put together a test class and methods
>
>> setClass("foo",representation(x="numeric"))
> [1] "foo"

This creates an *S4* class which HAS A 'slot' x that is of type numeric.

>> setMethod("plot","foo",function(x,y,...)boxplot(x,...))
> [1] "plot"
>> x <- rnorm(100)
>> class(x) <- "foo"

uh oh, this is creating an *S3* class that IS A numeric that has been  
given the class name "foo". The S3 class 'foo' and S4 class 'foo' are  
different. Note that you can make anything an S3 class

> z=letters
> class(z) <- "foo"

because S3 does not have any definition for what a 'foo' class is.

>> plot(x)

this 'works' but in my view should not -- the method is defined for  
the S4 class, not for the S3 class.

>> y <- new("foo",x=rnorm(50))

Here you're creating an instance of the S4 class. S4 insists that  
objects constructed as class 'foo' conform to the class specification

> new("foo", x=letters)
Error in validObject(.Object) :
   invalid class "foo" object: invalid object for slot "x" in class  
"foo": got cl
ass "character", should be or extend class "numeric"

>> class(y)
> [1] "foo"
>
>
> My question is:  should I have a "foo" function which sets the class
> to "foo", please?  or a sort of Validity checker, please?

I think normally one would have class "Foo" (capitalized; this is my  
style & convention, others are free to do as they please) with  
constructor


setClass("Foo", representation=representation(x="numeric"))

Foo <- function(x=numeric(), ...)
{
     new("Foo", x=x, ...)
}

I'd normally also define a validity method with setValidity("Foo",  
<etc>) if my class had constraints other than 'slot x has to be  
numeric' (which is checked by the default validity method). Dispatch  
on validity methods is different from on standard methods, as  
explained in ?setValidity.

Hope that helps,

Martin

> Thanks,
> Edna Bell
>
> ______________________________________________
> R-help at r-project.org 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.
>