Skip to content

parmat error from MASS?

4 messages · john polo, Brian Ripley, Greg Snow

john polo
# 
i'm in the midst of using R for a PVA/matrix project. my math skills are 
lousy and programming skills are worse. to deal with an error that i 
couldn't solve, i changed a parameter that used a new part of a routine 
in our PVA package's code:

else if (ES == "ParDraw") {

        require(MASS)


(previously, ES was "MatrixDraw" and MASS wasn't called for)
now, i get the following error:

Error in apply(parmat, 1, mean) : dim(X) must have a positive length


there is no code in the PVA package routines that has a line that says 
"apply(parmat, 1, mean)" or parmat or the format "(____, 1, mean)".  
that's why i think the error has something to do with MASS. i tried to 
update MASS, but had problems with that. i've searched for parmat. i 
looked through the MASS documentation for parmat. i don't know what it 
is or how to change whatever is affecting parmat.

i'm using RGui from version 2.6.0 on an XP machine.

sincerely,
john
Brian Ripley
# 
There is no such code in MASS, so please don't make random assignments of
blame.  A competent R author would have used rowMeans(parmat).

R has debugging tools (see 'Writing R Extensions') and you should use them 
to locate in which function the error is occurring.  E.g. traceback() 
would tell you.  Even then, the error is almost certainly in the data 
passed to the function and not in the function itself.

We don't know what the PVA package/project is, or who 'our' refers to -- 
despite the posting guide you have not given a signature block.
On Mon, 10 Mar 2008, john polo wrote:

            

      
      
      
    

        

      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      john polo
    

    

      
      #
    

  


  

      
thank you for your help.

        
Prof Brian Ripley wrote:

            

      
      
      
    

        
that is why i asked for help. i've been using R for 2 months now. i 
never claimed to be competent.

            

      
      
      
    

        
thank you for pointing me in the direction for error interpretation. i 
posted once before, about a month ago and asked if there was a way to 
learn how to interpret error messages. the only response was that i'd 
learn to recognize them after some time. you were correct, traceback 
indicated MASS was not at fault.

from a non-programmer, non-expert point of view, if i was going to try 
to locate a source for error interpretation (which i have spent many 
hours over the last 2 months doing. it's much easier to do that than 
come here and get helpful, if snarky, responses), i would have thought 
there would be information in the refman or intro or maybe installation 
and administration. as is obvious, i'm nowhere near writing packages for 
R. reading such documentation wasn't obvious for me.

            

      
      
      
    

        
i had intended to include a link to the PVA package, but it seems the 
author (Bruce Kendall @ UCSB) took it down. i didn't think the list 
wanted me to post all the code from the routines i have been using.
the posting guideline said:
"/Some/ consider it good manners to include a concise signature 
specifying affiliation"
my emphasis. if it's required, the guidelines should say that. had i a 
sig it would say University of South Florida, where i'm an undergrad in 
biology. i don't know if that explains much more about my lack of R or 
computer or math competence.

sincerely,
john
University of South Florida