[R] OOP and passing by value
Joris Meys
jorismeys at gmail.com
Wed Jun 9 16:00:33 CEST 2010
Thanks for the correction. Personally, I never really use that kind of
assignments, and the deparse(substitute()) construct I only use in
self-defined plotting functions in simple scripts to be run on series
of equivalent datasets by me, myself and I.
When writing a package with nested functions and the likes, there are
indeed far better solutions than this one. I am -obviously- not an
expert on those. And honestly, I personally don't like the idea of a
function changing my object. So in conclusion, I shouldn't even have
sent my last mail I guess...
Cheers
Joris
On Wed, Jun 9, 2010 at 3:12 PM, Henrik Bengtsson <hb at stat.berkeley.edu> wrote:
> Pass an object of, or containing an, environment. Then whenever you
> modify any object "inside" the environment, the changes will remain
> also when exiting from the function(s). This has been used by many
> for quite some time and is the standard way to do it, if you need this
> feature. See packages such as R.oo and proto for complete solutions.
>
> Using substitute(this) is not safe when doing nested calling, and
> assigning to a global environment is not something you really want to
> do.
>
> /Henrik
>
> On Wed, Jun 9, 2010 at 3:05 PM, Joris Meys <jorismeys at gmail.com> wrote:
>> In case you want the function setData to change the object itself
>> (which is often a dangerous idea!), you can use instead :
>>
>> setMethod("setData","test",
>> function(this,fcn,k){
>> Name <- deparse(substitute(this))
>> this at t <- as.numeric(seq(-k,k))/(2*k+1)
>> this at f <- sapply(this at t,FUN=fcn)
>> assign(Name,this,.GlobalEnv)
>> }
>> )
>>
>> #-----------------------------------------------------------
>> tst <- new("test")
>> fcn <- function(u){ sin(2*pi*u) }
>> setData(tst,fcn,5)
>> tst at t
>> [1] -0.4545455 -0.3636364 -0.2727273 -0.1818182 -0.0909091 0.0000000
>> 0.0909091 0.1818182 0.2727273 0.3636364 0.4545455
>>
>>
>> On Wed, Jun 9, 2010 at 2:36 PM, Joris Meys <jorismeys at gmail.com> wrote:
>>> by returning the object ?
>>>
>>> setMethod("setData","test",
>>> function(this,fcn,k){
>>> this at t <- as.numeric(seq(-k,k))/(2*k+1)
>>> this at f <- sapply(this at t,FUN=fcn) #changed!
>>> return(this) #changed!
>>> }
>>> )
>>>
>>> #-----------------------------------------------------------
>>>> tst <- new("test")
>>>> fcn <- function(u){ sin(2*pi*u) }
>>>> New <- setData(tst,fcn,5)
>>>> New at t
>>> [1] -0.4545455 -0.3636364 -0.2727273 -0.1818182 -0.0909091 0.0000000
>>> 0.0909091 0.1818182 0.2727273 0.3636364 0.4545455
>>>
>>>
>>> On Wed, Jun 9, 2010 at 2:28 PM, michael meyer <mjhmeyer at googlemail.com> wrote:
>>>> Greetings,
>>>>
>>>> I love the R system and am sincerely grateful for the great effort the
>>>> product and contributors
>>>> are delivering.
>>>>
>>>> My question is as follows:
>>>>
>>>> I am trying to use S4 style classes but cannot write functions that modify
>>>> an object
>>>> because paramter passing is by value.
>>>> For example I want to do this:
>>>>
>>>> setGeneric("setData", function(this,fcn,k){ standardGeneric("setData") })
>>>>
>>>> setClass(
>>>> "test",
>>>> representation(f="numeric", t="numeric")
>>>> )
>>>> setMethod("setData","test",
>>>> function(this,fcn,k){
>>>> this at t <- as.numeric(seq(-k,k))/(2*k+1)
>>>> this at f <- sapply(t,FUN=fcn)
>>>> }
>>>> )
>>>>
>>>> #-----------------------------------------------------------
>>>> tst <- new("test")
>>>> fcn <- function(u){ sin(2*pi*u) }
>>>> setData(tst,fcn,100)
>>>> tst at t # it's still empty because of pass by value
>>>>
>>>>
>>>> How can this be handled?
>>>>
>>>>
>>>> Many thanks,
>>>>
>>>> Michael
>>>>
>>>> [[alternative HTML version deleted]]
>>>>
>>>> ______________________________________________
>>>> R-help at r-project.org mailing list
>>>> https://stat.ethz.ch/mailman/listinfo/r-help
>>>> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
>>>> and provide commented, minimal, self-contained, reproducible code.
>>>>
>>>
>>>
>>>
>>> --
>>> Joris Meys
>>> Statistical consultant
>>>
>>> Ghent University
>>> Faculty of Bioscience Engineering
>>> Department of Applied mathematics, biometrics and process control
>>>
>>> tel : +32 9 264 59 87
>>> Joris.Meys at Ugent.be
>>> -------------------------------
>>> Disclaimer : http://helpdesk.ugent.be/e-maildisclaimer.php
>>>
>>
>>
>>
>> --
>> Joris Meys
>> Statistical consultant
>>
>> Ghent University
>> Faculty of Bioscience Engineering
>> Department of Applied mathematics, biometrics and process control
>>
>> tel : +32 9 264 59 87
>> Joris.Meys at Ugent.be
>> -------------------------------
>> Disclaimer : http://helpdesk.ugent.be/e-maildisclaimer.php
>>
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>>
>
--
Joris Meys
Statistical consultant
Ghent University
Faculty of Bioscience Engineering
Department of Applied mathematics, biometrics and process control
tel : +32 9 264 59 87
Joris.Meys at Ugent.be
-------------------------------
Disclaimer : http://helpdesk.ugent.be/e-maildisclaimer.php
More information about the R-help
mailing list