Roger Koenker wrote:
>
> I would like to add a class to the SparseM package. I have a class "matrix.csr"
> that describes a matrix in compressed sparse row format, now I would like a class
> matrix.diag.csr that describes such objects when they happen to be diagonal.
> The idea is that matrix.diag.csr objects should behave (later in life) exactly like
> matrix.csr objects, the distinction is only needed in order to clarify the
> situation when they are created. So I have:
>
> setClass("matrix.diag.csr","matrix.csr")
>
> My understanding (shakey, as it is) is that if I now try to construct a
> matrix.diag.csr
> object the checking of it should be just like the checking for a matrix.csr object,
> but
Well, hopefully. But there seems a slight problem in the validity
method.
You require a dimension slot of length 2, but the default object
generated from matrix.csr fails that:
R> library(SparseM)
[1] "SparseM library loaded"
R> new("matrix.csr")
Error in validObject(.Object) : Invalid "matrix.csr" object: invalid
dimension attribute
Enter a frame number, or 0 to exit
1:new("matrix.csr")
2:initialize(value, ...)
3:initialize(value, ...)
4:.local(.Object, ...)
5:validObject(.Object)
Selection: 5
Browse[1]> [EMAIL PROTECTED]
numeric(0)
(Your extension of the class uncovered the problem by generating the
default object, as the traceback below showed.)
John
>
> > new("matrix.csr",ra=1:3,ja=1:3,ia=1:4,dimension=as.integer(c(3,3)))
> An object of class "matrix.csr"
> Slot "ra":
> [1] 1 2 3
>
> Slot "ja":
> [1] 1 2 3
>
> Slot "ia":
> [1] 1 2 3 4
>
> Slot "dimension":
> [1] 3 3
>
> > new("matrix.diag.csr",ra=1:3,ja=1:3,ia=1:4,dimension=as.integer(c(3,3)))
> Error in validObject(.Object) : Invalid "matrix.csr" object: invalid dimension
> attribute
>
> the traceback() says:
>
> 16: stop(paste("Invalid \"", Class, "\" object: ", errors, sep = ""))
> 15: validObject(.Object)
> 14: .local(.Object, ...)
> 13: initialize(value, ...)
> 12: initialize(value, ...)
> 11: new("matrix.csr")
> 10: asMethod(object)
> 9: as(object, superClass)
> 8: validityMethod(as(object, superClass))
> 7: identical(x, TRUE)
> 6: anyStrings(validityMethod(as(object, superClass)))
> 5: validObject(.Object)
> 4: .local(.Object, ...)
> 3: initialize(value, ...)
> 2: initialize(value, ...)
> 1: new("matrix.diag.csr", ra = 1:3, ja = 1:3, ia = 1:4, dimension = as.integer(c(3,
> 3)))
>
> which might be revelatory to some of you, but leaves me in the dark. Any suggestions
> would be greatly appreciated. In case it is relevant
>
> setClass("matrix.csr",representation(ra="numeric",
> ja="integer",ia="integer", dimension="integer"),
> validity = function(object) {
> if(!length([EMAIL PROTECTED]) == 2 )
> return("invalid dimension attribute")
> else{
> nrow <- [EMAIL PROTECTED]
> ncol <- [EMAIL PROTECTED]
> }
> if(!(length([EMAIL PROTECTED]) ==length([EMAIL PROTECTED])))
> return("ra and ja don't have equal lengths")
> if(any([EMAIL PROTECTED] < 1) || any([EMAIL PROTECTED] > ncol))
> return("ja exceeds dim bounds")
> if(any([EMAIL PROTECTED] < 1))
> return("some elements of ia are <= 0")
> if(any(diff([EMAIL PROTECTED])<0))
> return("ia vector not monotone increasing")
> if([EMAIL PROTECTED]([EMAIL PROTECTED])] != length([EMAIL
> PROTECTED])+1)
> return("last element of ia doesn't conform")
> if(length([EMAIL PROTECTED]) != nrow+1)
> return("ia has wrong number of elments")
> if(length([EMAIL PROTECTED]) < 1 || length([EMAIL PROTECTED]) >
> nrow*ncol)
> return("ra has too few, or too many elements")
> TRUE})
> setMethod("initialize", "matrix.csr",
> function(.Object, ra = numeric(0), ja = integer(0),
> ia = integer(0),dimension = integer(0)) {
> [EMAIL PROTECTED] <- ra
> [EMAIL PROTECTED] <- ja
> [EMAIL PROTECTED] <- ia
> [EMAIL PROTECTED] <- dimension
> validObject(.Object)
> .Object
> })
>
> url: www.econ.uiuc.edu/~roger/my.html Roger Koenker
> email [EMAIL PROTECTED] Department of Economics
> vox: 217-333-4558 University of Illinois
> fax: 217-244-6678 Champaign, IL 61820
>
> ______________________________________________
> [EMAIL PROTECTED] mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
--
John M. Chambers [EMAIL PROTECTED]
Bell Labs, Lucent Technologies office: (908)582-2681
700 Mountain Avenue, Room 2C-282 fax: (908)582-3340
Murray Hill, NJ 07974 web: http://www.cs.bell-labs.com/~jmc
______________________________________________
[EMAIL PROTECTED] mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
