[R] Check the class of an object

[Martin Morgan]

On 07/23/2013 09:59 AM, Simon Zehnder wrote:
> Hi David,
>
> thanks for the reply. You are right. Using the %in% is more stable and I gonna change my code.

you said you were you were using S4 classes. S4 classes do not report vectors of 
length != 1, from ?class

      For objects which have a formal class, its name is returned by 'class'
      as a character vector of length one

so a first unit test could be

   stopifnot(length(class(myObject)) != 1L)


>
> When testing for a specific class using 'is' one has to start at the lowest heir and walk up the inheritance structure. Starting at the checks at the root will always give TRUE. Having a structure which is quite complicated let me move to the check I suggested in my first mail.
>
> Best
>
> Simon
>
> On Jul 23, 2013, at 6:15 PM, David Winsemius <dwinsemius at comcast.net> wrote:
>
>>
>> On Jul 23, 2013, at 5:36 AM, Simon Zehnder wrote:
>>
>>> Dear R-Users and R-Devels,
>>>
>>> I have large project based on S4 classes. While writing my unit tests I found out, that 'is' cannot test for a specific class, as also inherited classes can be treated as their super classes. I need to do checks for specific classes. What I do right now is sth. like
>>>
>>> if (class(myClass) == "firstClass") {
>>
>> I would think that you would need to use `%in%` instead.
>>
>> if( "firstClass" %in% class(myObject) ){
>>
>> Objects can have more than one class, so testing with "==" would fail in those instances.
>>
>>
>>>
>>> } else if (class(myClass) == "secondClass") {
>>>
>>> }
>>>
>>> Is this the usual way how classes are checked in R?
>>
>> Well, `inherits` IS the usual way.
>>
>>> I was expecting some specific method (and 'inherits' or 'extends' is not what I look for)...
>>>
>>>
>>> Best
>>>
>>> Simon
>>>
>>> 	[[alternative HTML version deleted]]
>>
>> Plain-text format is the recommended format for Rhelp
>>
>> --
>> David Winsemius
>> Alameda, CA, USA
>>
>
> ______________________________________________
> 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.
>


-- 
Computational Biology / Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N.
PO Box 19024 Seattle, WA 98109

Location: Arnold Building M1 B861
Phone: (206) 667-2793