> "PD" == Peter Dalgaard <[EMAIL PROTECTED]>
> on Fri, 16 Mar 2007 08:09:00 +0100 writes:
PD> Seth Falcon wrote:
>> [EMAIL PROTECTED] writes:
>>
>>> Investigating this new implementation I come across an issue in
>>> conjuntion with using S4 classes and methods. t
Hi,
I've noted that not all `predict' methods are hidden in the namespace:
> methods("predict")
[1] predict.ar*predict.Arima*
[3] predict.arima0*predict.glm
[5] predict.HoltWinters* predict.lm
[7] predict.loess* predict.mlm
[9] predict.nls*
One reason is that you should never call those methods directly but
only via the generic function predict(), and this is a way to prevent
such misuse/mishaps. You can always use getAnywhere() if you want
look at the code.
/H
On 3/16/07, Henric Nilsson (Public) <[EMAIL PROTECTED]> wrote:
> Hi,
>
The reason some are not hidden is historical: they were called explicitly
by (contributed) package code at the time namespaces were introduced.
It would be a good idea to revisit some of those decisions, but it is not
a very appealing way to spend one's time.
On Fri, 16 Mar 2007, Henric Nilsson
I believe your assertions are incorrect about both '=' and when copying is
done.
'=' is neither old nor deprecated and is syntactically different from
'<-', but where it means assignment, it invokes the same internal code as
'<-'.
R maintains a call-by-value illusion, but it does not actually
Den 2007-03-16 13:25, Prof Brian Ripley skrev:
> The reason some are not hidden is historical: they were called
> explicitly by (contributed) package code at the time namespaces were
> introduced.
Ah, I see.
In that case I'll just hide away as much as possible in the new package
I'm working o
Sometimes its useful to call a method on a class other than the method
indicated by the suffix of the method name and in that case one needs
to call the method directly rather than via the generic.
For an example, see:
http://tolstoy.newcastle.edu.au/R/e2/help/07/01/9383.html
On 3/16/07, Henric
This is off-topic, but since the discussion moved towards coding
style... Here are some comments on S4 style.
[EMAIL PROTECTED] writes:
> ## using S4 classes and methods
> setClass("fooBase",
> representation("VIRTUAL",
> width = "numeric",
>
cstrato <[EMAIL PROTECTED]> writes:
> I have a variation of this question:
> Is it possible to use the same function name, e.g. "myfunction" in both,
> an S4
> baseClass and derivedClass. The method "myfunction" in derivedCalss should
> extend the functionality defined in baseClass, analogously to
Is there a (C or R) function to invalidate the relevant parts of SEXP
/ data that are inaccessible? My problem is in tracking down
protection and other memory mismanagement bugs, where I have to rely
on luck to invalidate or overwrite data to trigger a detectable
error. This makes it hard to track
The 'Writing R Extensions' manual, configure.ac, and the comment at
the top of memory.c suggests that there are valgrind instrumentation
levels 0, 1, 2 but I think the actual code has four levels.
level 0 is no additional instrumentation
level 1 marks uninitialized numeric, logical, integer
Thank you Ei-ji,
That seems to have done it.
> .Machine$sizeof.pointer
[1] 8
So in total I've done the following to get it work:
Installed these readline rpms from the SLES10 media:
readline-devel-64bit-5.1-24.4
readline-devel-5.1-24.4
This is the configure command:
./configure CC="gcc -m64
The following redefines $ on "myclass" objects so that it looks up
the concatenation of the unevaluated second argument and 2
in the object and returns the result:
"$.myclass" <- function(x, name) x[[paste(substitute(name), 2, sep = "")]]
# test
myobject <- structure(list(x2 = 3), class = "myclas
Dear Seth
Thank you, maybe I still do not understand S4 methods.
I thought that the purpose of function "callNextMethod()" is to allow
some kind of inheritance. In the Bioconuductor packages which are
using S4, it is only used for method("initialize") but I thought, that
it can be used with every
Dear all,
Currently, I am trying to create a test package "testS4" using S4 classes,
which I am attaching.
Running R CMD check gives the following error:
coeurebooks-computer:/Volumes/CoreData/CRAN cs$ R CMD check
testS4_0.1.1.tar.gz
* checking for working latex ...sh: line 1: latex: command
your namespace probably contains:
useDynLib(testS4)
but you don't have any compiled code...
remove that line and everything will be fine.
btw, if you're playing with S4, you must import 'methods'
b
On Mar 16, 2007, at 4:55 PM, cstrato wrote:
> Dear all,
>
> Currently, I am trying to create a
Full_Name: Vladimir Obolonkin
Version: tested in 2.0 to 2.4.1
OS: linux, win, mac
Submission from: (NULL) (202.14.96.194)
{{ Subject shortened manually -- to pass anti-spam filters
Original-subject: Incorrect matrix of spearman correlations working \
large (24000 by 425 and 78 by 425 dat
Dear Benilton
Thank you, now my package works.
Do you have an explanation for my second problem:
* Installing *source* package 'mytest' ...
** R
** save image
Error in setMethod("export", "derivedClass", export.derivedClass) :
no existing definition for function "export"
Error: unable to
[EMAIL PROTECTED] wrote:
> Full_Name: Vladimir Obolonkin
> Version: tested in 2.0 to 2.4.1
> OS: linux, win, mac
> Submission from: (NULL) (202.14.96.194)
>
> {{ Subject shortened manually -- to pass anti-spam filters
>
>Original-subject: Incorrect matrix of spearman correlations working \
>
[EMAIL PROTECTED] wrote:
> Full_Name: Vladimir Obolonkin
> Version: tested in 2.0 to 2.4.1
> OS: linux, win, mac
> Submission from: (NULL) (202.14.96.194)
>
> {{ Subject shortened manually -- to pass anti-spam filters
>
>Original-subject: Incorrect matrix of spearman correlations working \
>
do you have a generic for 'export'?
On Mar 16, 2007, at 5:38 PM, cstrato wrote:
> Dear Benilton
>
> Thank you, now my package works.
>
> Do you have an explanation for my second problem:
>
> * Installing *source* package 'mytest' ...
> ** R
> ** save image
> Error in setMethod("export", "derivedC
cstrato <[EMAIL PROTECTED]> writes:
> Dear Seth
>
> Thank you, maybe I still do not understand S4 methods.
>
> I thought that the purpose of function "callNextMethod()" is to allow
> some kind of inheritance. In the Bioconuductor packages which are
> using S4, it is only used for method("initializ
Dear Seth
Thank you, I will further explain this feature.
Best regards
Christian
__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
Dear Benilton
Yes, but the error disappears only when I define it twice.
For my package I create three files:
1. MyClasses.R:
Here I define the classes only using
setClass("baseClass")
setClass("derivedClass")
2. methods.baseClass.R
Here I define the methods for baseClass
setGeneric("export", f
Well, my understanding is that if you "need" to define it (the
generic) twice you're doing something wrong.
Do you have a collate field in your description file? IMHO, the
"AllGeneric.R"/"AllClasses.R" approach is the way to go... keeping
your code organized is always a good thing, right?
b
I agree, but I hope that "AllGeneric.R" is not the only possibility.
BTW, what is a collate field?
Best regards
Christian
Benilton Carvalho wrote:
> Well, my understanding is that if you "need" to define it (the
> generic) twice you're doing something wrong.
>
> Do you have a collate field in yo
The Collate field gives the order that the files are going to be
concatenated. If they're concatenated using the "wrong" order, your
installation might fail (for example, the generic is defined after
the method).
A recommended source is the "Writing R Extensions"
http://cran.r-project.org/
Thank you, I have already read the R-exts manual, but the collate field
is optional.
Do you know a package which uses the Collate field so that I can study it?
Best regards
Christian
Benilton Carvalho wrote:
> The Collate field gives the order that the files are going to be
> concatenated. If t
Biobase itself does.
b
On Mar 16, 2007, at 7:43 PM, cstrato wrote:
> Thank you, I have already read the R-exts manual, but the collate
> field is optional.
> Do you know a package which uses the Collate field so that I can
> study it?
>
> Best regards
> Christian
>
> Benilton Carvalho wrote:
"Matthias Burger" <[EMAIL PROTECTED]> writes:
> Well, I understand I should have spent more time trimming my example
> to the minimal required code. Again this piece was copied & reduced.
No need to have done anything differently. I commented on the code
in hopes that it might be helpful for othe
I must have missed this, thank you.
Good night
Christian
Benilton Carvalho wrote:
> Biobase itself does.
> b
>
> On Mar 16, 2007, at 7:43 PM, cstrato wrote:
>
>> Thank you, I have already read the R-exts manual, but the collate
>> field is optional.
>> Do you know a package which uses the Collat
31 matches
Mail list logo
