The R Project SVN R

% File src/library/methods/man/setGeneric.Rd
% Part of the R package, http://www.R-project.org
% Copyright 1995-2007 R Core Development Team
% Distributed under GPL 2 or later

\name{setGeneric}
\alias{setGeneric}
\alias{setGroupGeneric}
\title{Define a New Generic Function}
\description{
  Create a new generic function of the given name, that is, a function
  that dispatches methods according to the classes of the arguments,
  from among the formal methods defined for this function.
}
\usage{
setGeneric(name, def= , group=list(), valueClass=character(),
           where= , package= , signature= , useAsDefault= ,
           genericFunction= , simpleInheritanceOnly = )

setGroupGeneric(name, def= , group=list(), valueClass=character(),
                knownMembers=list(), package= , where= )
}
\arguments{
  \item{name}{ The character string name of the generic function.
    The simplest (and recommended) call, \code{setGeneric(name)}, 
    looks for a function with this name and creates a corresponding
    generic function,  if the function found was not generic.
  }
  \item{def}{ An optional function object, defining the generic.
      Don't supply this argument if you want an existing non-generic
      function to supply the arguments. Do supply it if there is
    no current function of this name, or if you want the generic
    function to have different arguments. In that case, the formal arguments
    and default values for the generic are taken from \code{def}.  You
    can also supply this argument if you want the generic function to
    do something other than just dispatch methods.

    Note that \code{def} is \emph{not} the default method; use argument
    \code{useAsDefault} if you want to specify the default separately.
  }
  \item{group}{ Optionally, a character string giving the name of the group
    generic function to which this function belongs.    See
    \link{Methods} for details of group generic functions in method selection.
  }
  \item{valueClass}{ An optional character vector of one or more class
      names.  The value returned by the generic function must
    have (or extend) this class, or one of the classes; otherwise,
    an error is generated.
  }
  \item{package}{ The name of the package with which this function is
    associated.  Usually determined automatically (as the package
    containing the non-generic version if there is one, or else the
    package where this generic is to be saved).
  }
  \item{where}{ Where to store the resulting initial methods definition,
    and possibly the generic function; by
    default, stored into the top-level environment.
  }
  \item{signature}{
    Optionally, the vector of names, from among the formal arguments
    to the  function, that can
   appear in the signature of methods for this function, in calls to
   \code{\link{setMethod}}. If \dots is one of the
   formal arguments, it is treated specially. Starting with version
   2.8.0 of \R, \dots may be signature of the generic function.
   Methods will then be selected if their signature matches all the
   \dots arguments.  See the documentation for topic
   \link{dotsMethods} for details.  In the present version, it is not
   possible to mix \dots and other arguments in the signature (this
   restriction may be lifted in later versions).

   By default, the signature is inferred from the implicit generic
   function corresponding to a  non-generic function.  If no implicit generic function has been defined,  the default is all the
   formal arguments except \dots, in the order they appear in the
   function definition. In the case that
   \dots is the only formal argument, that is also the default signature.  To use \dots as the signature
   in a function that has any other arguments, you must supply the
   signature argument explicitly.  See the \dQuote{Implicit
      Generic} section below for more details.
  }
  \item{useAsDefault}{
    Override the usual choice of default argument (an existing
    non-generic function or no default if there is no such function).
    Argument \code{useAsDefault} can be supplied,
    either as a function to use for the default, or as a logical value.
    \code{FALSE} says not to have a default method at all, so that an
    error occurs if there is not an explicit or inherited method for a
    call.
    \code{TRUE} says to use the existing function as default,
    unconditionally (hardly ever needed as an explicit argument).
    See the section on details.
  }
  \item{simpleInheritanceOnly}{
      Supply this argument as \code{TRUE} to require that methods
      selected be inherited through simple inheritance only; that is,
      from superclasses specified in the \code{contains=} argument to
      \code{\link{setClass}}, or by simple inheritance to a class
      union or other virtual class.  Generic functions should require
      simple inheritance if they need to be assured that they get the
      complete original object, not one that has been transformed.
      Examples of functions requiring simple inheritance are
      \code{\link{initialize}}, because by definition it must return
      an object from the same class as its argument, and
      \code{\link{show}}, because it claims to give a full description
      of the object provided as its argument.
    }
  \item{genericFunction}{ Don't use; for (possible) internal use only.
  }

  \item{knownMembers}{
    (For \code{setGroupGeneric} only.)  The names of functions that are
    known to be members of this group.  This information is used to
    reset cached definitions of the member generics when information
    about the group generic is changed.
  }
}
\section{Basic Use}{
  The \code{setGeneric} function is called to initialize a generic
  function as
  preparation for defining some methods for that function.

  The simplest and most common situation is that \code{name} is already
  an ordinary non-generic non-primitive function, and you now want to turn this
  function into a generic.  In this case you will most often supply only
  \code{name}, for example:

    \code{setGeneric("colSums")}

There must be an existing function of this name, on some  attached package (in this case package \code{"base"}).
A generic version of this function will be created in the current package (or in the global environment if the call to \code{setGeneric()} is from an ordinary source file or is entered on the command line).
 The existing function becomes the default method, and
the package slot of the new generic function is set to the location of the original function (\code{"base"} in the example).  It's an important feature that the same generic function definition is created each time, depending in the example only on the definition of \code{print} and where it is found.
The \code{signature} of the generic function, defining which of the formal arguments can be used in specifying methods, is set by default to all the formal arguments except \dots.

Note that calling \code{setGeneric()} in this form is not strictly necessary before calling \code{setMethod()} for the same function.  If the function specified in the call to \code{setMethod} is not generic, \code{setMethod} will execute the call to \code{setGeneric} itself.  Declaring explicitly that you want the function to be generic can be considered better programming style; the only difference in the result, however, is that not doing so produces a message noting the creation of the generic function.

  You cannot (and never need to) create an explicit generic version of the
  primitive functions in the base package.  Those which can be treated
  as generic functions have methods selected and
  dispatched from the internal C code, to satisfy concerns for  efficiency, and the others cannot be made
  generic.  See the section on Primitive Functions below.

The description above is the effect when the package that owns the non-generic function has not created an implicit generic version.  Otherwise, it is this implicit generic function that is used. See the section on Implicit Generic Functions below.  Either way, the essential result is that the \emph{same} version of the  generic function will be created each time.

 The second common use of \code{setGeneric()}  is to create a new generic
  function, unrelated to any existing function, and frequently having no default method.  In this case, you
  need to supply a skeleton of the function definition, to define the
  arguments for the function.  The body of a generic function is
  usually a standard form, \code{standardGeneric(name)} where
  \code{name} is the quoted name of the generic function.
  When calling \code{setGeneric} in this form, you would normally
  supply the \code{def} argument as a function of this form.  See the second and third examples below.

The \code{useAsDefault} argument controls the default method for the new generic.
 If not
  told otherwise, \code{setGeneric} will try to find a non-generic
  version of the function to use as a default. 
So, if you do have a suitable default method, it is often simpler to first set this up as a non-generic function, and then use the one-argument call to \code{setGeneric} at the beginning of this section.   See the first example in the Examples section below.

 If you \emph{don't} want the existing function to be taken as default, supply the argument \code{useAsDefault}.  That argument
  can be the function you want to be the default method, or
  \code{FALSE} to force no default (i.e., to cause an error if
  there is no direct or inherited method selected for a call to the function).
}
\section{Details}{
If you want to change the behavior of an existing function (typically, one in another package) when you create a generic version, you must supply arguments to \code{setGeneric} correspondingly.
Whatever changes are made, the new generic function will be assigned with a package slot set to the \emph{current} package, not the one in which the non-generic version of the function is found.
This step is required because the version you are creating is no longer the same as that implied by the function in the other package.
A message will be printed to indicate that this has taken place and noting one of the differences between the two functions.


  The body of a generic function usually does nothing
  except for dispatching methods by a call to \code{standardGeneric}.
  Under some circumstances you might just want to do some additional
  computation in the generic function itself.  As long as your
  function eventually calls \code{standardGeneric} that is permissible
  (though perhaps not a good idea, in that it may make the behavior of
  your function less easy to understand).  If your explicit
  definition of the generic function
  does \emph{not} call \code{standardGeneric} you are in trouble,
  because none of the methods for the function will ever be
  dispatched.

  By default, the generic function can return any object.  If
  \code{valueClass} is supplied, it should be a vector of class names;
  the value returned by a method is then required to satisfy
  \code{is(object, Class)} for one of the specified classes.  An empty
  (i.e., zero length) vector of classes means anything is allowed.  Note
  that more complicated requirements on the result can be specified
  explicitly, by defining a non-standard generic function.

  The \code{setGroupGeneric} function behaves like \code{setGeneric}
  except that it constructs a group generic function, differing in two
  ways from an ordinary generic function.  First, this function cannot
  be called directly, and the body of the function created will contain
  a stop call with this information.  Second, the group generic
  function contains information about the known members of the group,
  used to keep the members up to date when the group definition
  changes, through changes in the search list or direct specification
  of methods, etc.
}
\section{Implicit Generic Functions}{

Saying that a non-generic function \dQuote{is converted to a generic}
is more precisely state that the function is converted to the
corresponding \emph{implicit} generic function.
If no special action has been taken, any function corresponds
implicitly to a generic function with the same arguments, in which
 all arguments
    other than \dots can be used.  The signature of this generic
    function is the vector of formal arguments, in order, except for
    \dots.

The source code for a package can define an implicit generic function
version of any function in that package (see \link{implicitGeneric}
for the mechanism).  You can not, generally, define an implicit
generic function in someone else's package. The usual reason for
defining an implicit generic is to prevent certain arguments from
appearing in the signature, which you must do if you want the
arguments to be used literally or if you want to enforce lazy
evaluation for any reason.  An implicit generic can also contain some
methods that you want to be predefined; in fact, the implicit generic
can be any generic version of the non-generic function.  The implicit
generic mechanism can also be used to prohibit a generic version (see \code{\link{prohibitGeneric}}).

Whether defined or inferred automatically, the implicit generic will
be compared with the generic function that \code{setGeneric}
creates, when the implicit generic is in another package.  If the two
functions are identical, then the \code{package} slot of the created
generic will have the name of the package containing the implicit
generic.  Otherwise, the slot will be the name of the package in which
the generic is assigned.

The purpose of this rule is to ensure that all methods defined for a
particular combination of generic function and package names
correspond to a single, consistent version of the generic
function.  Calling \code{setGeneric} with only \code{name} and
possibly \code{package} as arguments guarantees getting the implicit
generic version, if one exists.

Including any of the other arguments can force a new, local version of
the generic function.  If you don't want to create a new version,
don't use the extra arguments. 
}
\section{Generic Functions and Primitive Functions}{
  A number of the basic \R functions are specially implemented as
  primitive functions, to be evaluated directly in the underlying C
  code rather than by evaluating an \R language definition.  Most have
  implicit generics (see \code{\link{implicitGeneric}}), and become
  generic as soon as methods (including group methods) are defined on
  them.  Others cannot be made generic.

  Even when methods are defined for such functions, the generic
  version is not visible on the search list, in order that the C
  version continues to be called.  Method selection will be initiated
  in the C code.  Note, however, that the result is to restrict
  methods for primitive functions to signatures in which at least one
  of the classes in the signature is a formal S4 class.

  To see the generic version of a primitive function, use
  \code{\link{getGeneric}(name)}.  The function
  \code{\link{isGeneric}} will tell you whether methods are defined
  for the function in the current session.

}
\value{
  The \code{setGeneric} function exists for its side effect: saving the
  generic function to allow methods to be specified later.  It returns
  \code{name}.
}
\references{
 Chambers, John M. (2008)
 \emph{Software for Data Analysis: Programming with R}
  Springer.  (For the R version.)

 Chambers, John M. (1998)
 \emph{Programming with Data}
 Springer (For the original S4 version.)
}
\examples{
\dontshow{
setClass("track", representation(x="numeric", y="numeric"))
}

## create a new generic function, with a default method
props <- function(object) attributes(object)
setGeneric("props")

## A new generic function with no default method
setGeneric("increment",
  function(object, step, ...)
    standardGeneric("increment")
)


###   A non-standard generic function.  It insists that the methods
###   return a non-empty character vector (a stronger requirement than
###    valueClass = "character" in the call to setGeneric)

setGeneric("authorNames",
    function(text) {
      value <- standardGeneric("authorNames")
      if(!(is(value, "character") && any(nchar(value)>0)))
        stop("authorNames methods must return non-empty strings")
      value
      })

\dontshow{
setMethod("authorNames", "character", function(text)text)

stopifnot(identical(authorNames(c("me", "you")), c("me", "you")),
          is(trySilent(authorNames(character())), "try-error"), # empty value
          is(trySilent(authorNames(NULL)), "try-error"))        # no default method
}

## An example of group generic methods, using the class
## "track"; see the documentation of setClass for its definition

## define a method for the Arith group

setMethod("Arith", c("track", "numeric"),
 function(e1, e2) {
  e1@y <- callGeneric(e1@y , e2)
  e1
})

setMethod("Arith", c("numeric", "track"),
 function(e1, e2) {
  e2@y <- callGeneric(e1, e2@y)
  e2
})

## now arithmetic operators  will dispatch methods:

t1 <- new("track", x=1:10, y=sort(stats::rnorm(10)))

t1 - 100
1/t1

\dontshow{
removeGeneric("authorNames")
removeClass("track")
removeMethods("Arith")
removeGeneric("props")
removeGeneric("increment")
}
}
\seealso{
  \code{\link{Methods}} and the links there for a general discussion,
  \code{\link{dotsMethods}} for methods that dispatch on
  \dQuote{\dots}, and \code{\link{setMethod}} for method definitions.
}
\keyword{ programming }
\keyword{ methods }