Skip to content

weird error with a lazyload .RData file in a package

5 messages · Michael Friendly, Uwe Ligges, Brian Ripley +1 more

Michael Friendly
# 
I added a new data file, NLSY.RData, to a package that uses LazyData: Yes
It passed R CRAN check and R CMD install worked w/o significant complaints.

* installing to library 'C:/R/R-2.15.2/library'
* installing *source* package 'heplots' ...
** R
** data
**  moving datasets to lazyload DB
** demo
** inst
** preparing package for lazy loading
Warning: package 'nnet' was built under R version 2.15.3
** help
  ...

Yet, when I try to use it, I get something incomprehensible:

 > library(heplots)
 > str(NLSY)
Error in str(NLSY) : internal error -3 in R_decompress1
In addition: Warning message:
In str(NLSY) : restarting interrupted promise evaluation
 >

But it works if I load directly from the file I imported into the package:

 > load("NLSY.RData")
 > str(NLSY)
'data.frame':   243 obs. of  6 variables:
  $ math    : num  50 28.6 50 32.1 21.4 ...
  $ read    : num  45.2 28.6 53.6 34.5 22.6 ...
  $ antisoc : int  4 0 2 0 0 1 0 1 1 4 ...
  $ hyperact: int  3 0 2 2 2 0 1 4 3 5 ...
  $ income  : num  52.52 42.6 50 6.08 7.41 ...
  $ educ    : int  14 12 12 12 14 12 12 12 12 9 ...

What could cause this?
Uwe Ligges
#
On 29.03.2013 15:24, Michael Friendly wrote:
No idea. Can you make the package source (incl that Rdata file) available?

Best,
Uwe Ligges
Brian Ripley
#
On 29/03/2013 14:24, Michael Friendly wrote:
That usually means that the LazyData database has been corrupted (and 
most likely was not created properly).  As to why: we need the 
reproducible example the posting guide asked you to provide.

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Peter Dalgaard
    

    

      
      #
    

  


  

      
On Mar 29, 2013, at 17:22 , Prof Brian Ripley wrote:

        

            

      
      
      
    

        
I had the same symptom a couple of days ago. It went away after quitting and restarting the R session. I kind of put it down to leaving an R session running while updating R (Yeah, I know: Still nothing reproducible...). The warning about nnet from 2.15.3 while installing to R-2.15.2/library suggests that something similar could be afoot in this case.

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Michael Friendly
    

    

      
      #
    

  


  

      
On 3/29/2013 12:05 PM, Uwe Ligges wrote:

            

      
      
      
    

        
It must have been something in my old session.  Starting a fresh one, it 
went away.