Skip to content

Problem with S3 method dispatch and NAMESPACE

4 messages · Peter Dalgaard, Frank E Harrell Jr

#
I have updated the rms package to extensively use NAMESPACE.  I cannot get
certain S3 methods to dispatch.  For example I have in NAMESPACE

S3method(anova, rms)
S3method(latex, anova.rms)

anova.rms produces an object of class "anova.rms" and there is a
latex.anova.rms function in rms.  But when I do latex(anova(fit)) I get an
invocation of latex.default.

I have tried using "anova.rms" and `anova.rms` in S3method() to no avail.

Any help appreciated.  I'm using R 2.15.3
Frank



-----
Frank Harrell
Department of Biostatistics, Vanderbilt University
--
View this message in context: http://r.789695.n4.nabble.com/Problem-with-S3-method-dispatch-and-NAMESPACE-tp4665179.html
Sent from the R devel mailing list archive at Nabble.com.
#
I found that package quantreg has created a new generic for latex() [I wish
it hadn't; this has been a generic in Hmisc for almost 2 decades].  When I
require(quantreg) after loading rms, latex(anova.rms object) dispatches
latex.default, but everything is fine if I don't load quantreg.  rms has
import(Hmisc) in NAMESPACE and is loaded before quantreg, hence the
conflict.  How do I make the generic from Hmisc take precedence?
Thanks
Frank

Frank Harrell wrote
-----
Frank Harrell
Department of Biostatistics, Vanderbilt University
--
View this message in context: http://r.789695.n4.nabble.com/Problem-with-S3-method-dispatch-and-NAMESPACE-tp4665179p4665216.html
Sent from the R devel mailing list archive at Nabble.com.
#
On Apr 24, 2013, at 15:59 , Frank Harrell wrote:

            
library(quantreg, pos=x) with suitably large x is one idea.

  
    
#
Thank you very much Peter.  That did the trick.
Frank

Peter Dalgaard-2 wrote

            

            
-----
Frank Harrell
Department of Biostatistics, Vanderbilt University
--
View this message in context: http://r.789695.n4.nabble.com/Problem-with-S3-method-dispatch-and-NAMESPACE-tp4665179p4665226.html
Sent from the R devel mailing list archive at Nabble.com.