The R Project SVN R

\name{validObject}
\alias{validObject}
\alias{setValidity}
\title{ Test the Validity of an Object }
\description{
  The validity of \code{object} related to its class definition is
  tested.  If the object is valid, \code{TRUE} is returned; otherwise,
  either a vector of strings describing validity failures is returned,
  or an error is generated (according to whether \code{test} is
  \code{TRUE}).  Optionally, all slots in the object can also be validated.

  The function \code{setValidity} sets the validity method of a class
  (but more normally, this method will be supplied as the
  \code{validity} argument to \code{\link{setClass}}).  The method
  should be a function of one object that returns \code{TRUE} or a
  description of the non-validity.
}
\usage{
validObject(object, test = FALSE, complete = FALSE)

setValidity(Class, method, where = topenv(parent.frame()) )
}
\arguments{
  \item{object}{ any object, but not much will happen unless the
    object's class has a formal definition.}
  \item{test}{logical; if \code{TRUE} and validity fails, the
    function returns a vector of strings describing the problems.  If
    \code{test} is \code{FALSE} (the default) validity failure generates
    an error.}
  \item{complete}{logical; if \code{TRUE}, validity methods will be
    applied recursively to any of the slots that have such methods.}
  \item{Class}{the name or class definition of the class whose validity
    method is to be set.}
  \item{method}{a validity method;  that is, either \code{NULL} or a
    function of one argument (\code{object}).  Like
    \code{validObject}, the function should return \code{TRUE} if the
    object is valid, and one or more descriptive strings if any problems
    are found.  Unlike \code{validObject}, it should never generate an
    error.
  }
  \item{where}{the modified class definition will be stored in this
    environment.}

  Note that validity methods do not have to check validity of
  superclasses: the logic of \code{validObject} ensures these tests are
  done once only.  As a consequence, if one validity method wants to use
  another, it should extract and call the method from the other
  definition of the other class by calling \code{\link{getValidity}}: it
  should \emph{not} call \code{validObject}.
}
\details{
  Validity testing takes place \dQuote{bottom up}: Optionally, if
  \code{complete=TRUE}, the validity of the object's slots, if any, is
  tested.  Then, in all cases, for each of the classes that this class
  extends (the \dQuote{superclasses}), the explicit validity method of
  that class is called, if one exists.  Finally, the validity method of
  \code{object}'s class is called, if there is one.

  Testing generally stops at the first stage of finding an error, except
  that all the slots will be examined even if a slot has failed its
  validity test.

  The standard validity test (with \code{complete=FALSE}) is applied
  when an object is created via \code{\link{new}} with any optional
  arguments (without the extra arguments the result is just the class
  prototype object).

  An attempt is made to fix up the definition of a validity method if
  its argument is not \code{object}.
}
\value{
  \code{validObject} returns \code{TRUE} if the object is valid.
  Otherwise a vector of strings describing problems found, except that
  if \code{test} is \code{FALSE}, validity failure generates an error,
  with the corresponding strings in the error message.
}
\references{
  The \R package \pkg{methods} implements, with a few exceptions, the
  programming interface for classes and methods in the book
  \emph{Programming with Data} (John M. Chambers, Springer, 1998), in
  particular sections 1.6, 2.7, 2.8, and chapters 7 and 8.

  While the programming interface for the \pkg{methods} package follows
  the reference, the R software is an original implementation, so
  details in the reference that reflect the S4 implementation may appear
  differently in R.  Also, there are extensions to the programming
  interface developed more recently than the reference.  For a
  discussion of details see \code{?\link{Methods}}
  and the links from that documentation.
}
\seealso{ \code{\link{setClass}}. }

\examples{
setClass("track",
          representation(x="numeric", y = "numeric"))
t1 <- new("track", x=1:10, y=sort(rnorm(10)))
## A valid "track" object has the same number of x, y values
validTrackObject <- function(object) {
    if(length(object@x) == length(object@y)) TRUE
    else paste("Unequal x,y lengths: ", length(object@x), ", ",
               length(object@y), sep="")
}
## assign the function as the validity method for the class
setValidity("track", validTrackObject)
## t1 should be a valid "track" object
validObject(t1)
## Now we do something bad
t2 <- t1
t2@x <- 1:20
## This should generate an error
\dontrun{try(validObject(t2))}
\dontshow{stopifnot(is(try(validObject(t2)), "try-error"))}

setClass("trackCurve",
         representation("track", smooth = "numeric"))

## all superclass validity methods are used when validObject
## is called from initialize() with arguments, so this fails
\dontrun{trynew("trackCurve", t2)}
\dontshow{stopifnot(is(try(new("trackCurve", t2)), "try-error"))}

setClass("twoTrack", representation(tr1 = "track", tr2 ="track"))

## validity tests are not applied recursively by default,
## so this object is created (invalidly)
tT  <- new("twoTrack", tr2 = t2)

## A stricter test detects the problem
\dontrun{try(validObject(tT, complete = TRUE))}
\dontshow{stopifnot(is(try(validObject(tT, complete = TRUE)), "try-error"))}
}
\keyword{programming}
\keyword{classes}