Group Generic Functions

Description

Group generic functions can be defined with either S3 and S4 methods (with different groups). Methods are defined for the group of functions as a whole.

A method defined for an individual member of the group takes precedence over a method defined for the group as a whole.

When package methods is attached there are objects visible with the names of the group generics: these functions should never be called directly (a suitable error message will result if they are).

Usage

## S3 methods have prototypes:
Math(x, ...)
Ops(e1, e2)
Summary(x, ...)
Complex(z)

## S4 methods have prototypes:
Arith(e1, e2)
Compare(e1, e2)
Ops(e1, e2)
Math(x)
Math2(x, digits)
Summary(x, ..., na.rm = FALSE)
Complex(z)

Arguments

S3 Group Dispatching

There are four groups for which S3 methods can be written, namely the "Math", "Ops", "Summary" and "Complex" groups. These are not R objects, but methods can be supplied for them and base R contains factor, data.frame and difftime methods for the first three groups. (There are also a ordered method for Ops, POSIXt methods for Math and Ops, as well as a ts method for Ops in package stats.)

1. Group "Math":
   • abs, sign, sqrt,
     floor, ceiling, trunc,
     round, signif
   • exp, log,
     cos, sin, tan,
     acos, asin, atan

     cosh, sinh, tanh,
     acosh, asinh, atanh

   • lgamma, gamma, gammaCody,
     digamma, trigamma
   • cumsum, cumprod, cummax, cummin
2. Group "Ops":
   • "+", "-", "*", "/", "^", "%%", "%/%"
   • "&", "|", "!"
   • "==", "!=", "<", "<=", ">=", ">"
3. Group "Summary":
   • all, any
   • sum, prod
   • min, max
   • range
4. Group Complex:
   • Arg, Conj, Im, Mod, Re

Note that a method will used for either one of these groups or one of its members only if it corresponds to a "class" attribute, as the internal code dispatches on oldClass and not on class. This is for efficiency: having to dispatch on, say, Ops.integer would be too slow.

The number of arguments supplied for "Math" group generic methods is not checked prior to dispatch. (Most have default methods expecting one argument, but three expect two.)

S4 Group Dispatching

When package methods is attached (which it is by default), formal (S4) methods can be defined for groups.

The functions belonging to the various groups are as follows:

Arith

"+", "-", "*", "^", "%%", "%/%", "/"

Compare

"==", ">", "<", "!=", "<=", ">="

Ops

"Arith", "Compare"

Math

"log", "sqrt", "log10", "cumprod", "abs", "acos", "acosh", "asin", "asinh", "atan", "atanh", "ceiling", "cos", "cosh", "cumsum", "exp", "floor", "gamma", "lgamma", "sin", "sinh", "tan", "tanh", "trunc"

Math2

"round", "signif"

Summary

"max", "min", "range", "prod", "sum", "any", "all"

Complex

"Arg", "Conj", "Im", "Mod", "Re"

Functions with the group names exist in the methods package but should not be called directly.

All the functions in these groups (other than the group generics themselves) are basic functions in R. They are not by default S4 generic functions, and many of them are defined as primitives, meaning that they do not have formal arguments. However, you can still define formal methods for them. The effect of doing so is to create an S4 generic function with the appropriate arguments, in the environment where the method definition is to be stored. It all works more or less as you might expect, admittedly via a bit of trickery in the background.

Note: currently those members which are not primitive functions must have been converted to S4 generic functions (preferably before setting an S4 group generic method) as it only sets methods for known S4 generics. This can be done by a call to setGeneric, for example setGeneric("round", group="Math2").

References

Appendix A, Classes and Methods of
Chambers, J. M. and Hastie, T. J. eds (1992) Statistical Models in S. Wadsworth & Brooks/Cole.

Chambers, J. M. (1998) Programming with Data. Springer, pp. 352–4.

See Also

methods for methods of non-Internal generic functions.

Examples

methods("Math")
methods("Ops")
methods("Summary")

d.fr <- data.frame(x=1:9, y=rnorm(9))
data.class(1 + d.fr) == "data.frame" ##-- add to d.f. ...

if(.isMethodsDispatchOn()) {# package "methods" is attached or loaded
  setClass("testComplex", representation(zz = "complex"))
  ## method for whole group "Complex"
  setMethod("Complex", "testComplex",
            function(z) c("groupMethod", callGeneric(z@zz)))
  ## exception for Arg() :
  setMethod("Arg", "testComplex",
            function(z) c("ArgMethod", Arg(z@zz)))
  z1 <- 1+2i
  z2 <- new("testComplex", zz = z1)
  stopifnot(identical(Mod(z2), c("groupMethod", Mod(z1))))
  stopifnot(identical(Arg(z2), c("ArgMethod", Arg(z1))))
}