[R] restrictions for attribute access <Watchdog: Virus checked>

[Prof Brian Ripley]

On Wed, 2 Apr 2008, Henrique Dallazuanna wrote:

> Work's for me:
>
> summary(lmer(Reaction ~ Days + (Days|Subject), sleepstudy))@AICtab

I think the issue is that in his package he is not seeing the summary S4 
generic from Matrix (which is the one lme4 sets methods on).  The package 
needs to import summary() from Matrix (and not lme4).

Try

> library(lme4)
...
> find("summary")
[1] "package:Matrix" "package:base"

and remember that base namespace comes before the search path in a package 
with a NAMESPACE.

Things get even more complicated if other S4-using packages are involved.

>
>
> On Wed, Apr 2, 2008 at 9:36 AM, Benjamin Polak <benjamin.polak at unisg.ch>
> wrote:
>
>> Thanks, but it doesn't work, I get the error message:
>>
>> 'no slot of name "AICtab" for this object of class "table"'
>>
>>
>> On Wednesday, 2. April 2008 13:34, Henrique Dallazuanna wrote:
>>> You've tried:
>>>
>>> tab <- smry at AICtab
>>>
>>> ?
>>>
>>> On Wed, Apr 2, 2008 at 7:59 AM, Benjamin Polak
>> <benjamin.polak at unisg.ch> wrote:
>>>> Hello everybody,
>>>>
>>>>  I have a problem with accessing class attributes. I think it
>>>> might be due to namespace access restrictions or something
>>>> similar, but I'm unable to figure it out, maybe someone knows the
>>>> trick how to do it ...
>>>>
>>>>  I'm trying to extract some information from the summary object
>>>> for a model fitted using the lmer function from the lme4 package:
>>>>
>>>>   smry <- summary(lmer(foo ~ bar ...))
>>>>   tab <- attr(smry, "AICtab")
>>>>   tab$AIC
>>>>
>>>>  When I run this code as a regular R script everything works
>>>> fine, tab$AIC has the desired value. But when I try to run this
>>>> code from within a package I wrote, tab$AIC is NULL.
>>>>
>>>>  I've tried adding "import(lme4)", "importClassesFrom(lme4)"
>>>>  and "importMethodsFrom(lme4)" into the NAMESPACE file of my
>>>> package, but it didn't help, tab$AIC is still NULL.
>>>>
>>>>  Thanks for any kind of hint!
>>>>
>>>>  Best Regards,
>>>>  Ben


-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595