[R] unable to use functions require DLL from package

David Winsemius dwinsemius at comcast.net
Sat Aug 16 03:24:01 CEST 2014


On Aug 15, 2014, at 2:56 PM, Charlotte de Vries wrote:

> David Winsemius <dwinsemius <at> comcast.net> writes:
> 
>> 
>> 
>> On Aug 15, 2014, at 8:06 AM, Charlotte de Vries wrote:
>> 
>>> Hey there!
>>> 
>>> I'm having problems with the same code, but I get a different error: 
>> 
>> This is apparently yet another example demonstrating why the Posting Guide 
> suggests that you include the
>> text of any earlier posting to which you want us to consider. At the 
> moment I see no posting that has this subject.
>> 
>>> 
>>> Error in .C("spline_coef", method = as.integer(method), n = n, x = 
>>> as.double(x),  : 
>>> "spline_coef" not available for .C() for package "stats"
>>> 
>>> 
>>> I'm using R3.1.0 on windows 8 and I've never used R before, so I might 
> have 
>>> made some terrible newby error (I have programmed quite a bit before, 
> but in C 
>>> and Matlab). 
>>> 
>>> Thank you!
>>> 
>>> Best,
>> 
>> David Winsemius
>> Alameda, CA, USA
>> 
>> 
> 
> Hi David, 
> 
> Thank you for answering. The original message I replied to was this message 
> (http://comments.gmane.org/gmane.comp.lang.r.general/113245):
> 
> **********************************************

Well, that explains why I don't have a copy on my machine. It's dated 6 May 18:15 2008


> Hi all, 
> 
> I have issues using some basic functions in R such as these ones : 

pp.test is not a "basic R function". I get this:

> ?pp.test
No documentation for ‘pp.test’ in specified packages and libraries:
you could try ‘??pp.test’

You are asked to include the package name for non-base R functions.

> 
>> pp.test(R) (where is a vector of returns)
> Error in .C("R_approx", as.double(x), as.double(y), as.integer(nx), xout =
> as.double(xout),  : 
>  C symbol name "R_approx" not in DLL for package "base"
> 
>> boxcox(reg,plotit=T)  (where reg is an lm object)
> Error in .C("spline_coef", method = as.integer(method), n = as.integer(nx), 
> : 
>  C symbol name "spline_coef" not in DLL for package "base"

Well I do have the MASS package from which that function was loaded. 

I get no error with the example in the help page:

 boxcox(Days+1 ~ Eth*Sex*Age*Lrn, data = quine,
         lambda = seq(-0.05, 0.45, len = 20), plotit=TRUE)

... and you are not providing a reproducible example that provokes the error, nor are you providing the version numbers of R or MASS.


> 
> as I do miss some symbol names. 
> 
> How can I overcome this serious problem ? 
> ***********************************
> 
> But someone else reported a very similar problem (oddly enough using the 
> same function spline_coef but this time from the package stats rather than 
> base (http://comments.gmane.org/gmane.comp.lang.r.general/115420): 

Which was in turn dated: 2 Jun 06:04 2008


> ************************************
> Subject: Missing "spline_coef" DLL and Rob Hyndmans monotonic	interpolator
> 
> Hello R help
> 
> I have been trying to use Rob Hyndman's monotonically increasing spline
> function.  But like another user or two seem have a problem with a
> missing DLL (namely "spline_coef").  None of the previous help postings
> seemed to have any solutions to this problem.  As per a Ripley
> suggestion I have deleted all previous versions of R and reinstalled R
> 2.7.0 and the problem persists.
> 
> Thanks
> 
> Paul.
> 
> x <- seq(0,4,l=20) 
> 
> y <- sort(rnorm(20)) 
> 
> plot(x,y) 
> lines(spline(x, y, n = 201), col = 2) # Not necessarily monotonic
> lines(cm.spline(x, y, n = 201), col = 3)

I get no error after:

library(demography)

# And then running that code.

I have an almost up-to-date version of R running on a mac (and at the moment I have way too many packages loaded:

> sessionInfo()
R version 3.1.0 Patched (2014-04-21 r65431)
Platform: x86_64-apple-darwin10.8.0 (64-bit)

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] grDevices datasets  splines   graphics  utils     stats    
[7] grid      methods   base     

other attached packages:
 [1] demography_1.17  ftsa_3.9         rainbow_3.2     
 [4] pcaPP_1.9-49     forecast_5.4     timeDate_3010.98
 [7] zoo_1.7-11       XML_3.98-1.1     RCurl_1.95-4.3  
[10] bitops_1.0-6     subplex_1.1-4    MASS_7.3-31     
[13] cobs_1.2-2       quantreg_5.05    gplots_2.13.0   
[16] reshape2_1.2.2   ggplot2_0.9.3.1  multcomp_1.3-1  
[19] TH.data_1.0-3    mvtnorm_0.9-9999 data.table_1.9.2
[22] muhaz_1.2.5      downloader_0.3   RJSONIO_1.3-0   
[25] plot3D_1.0-1     rms_4.2-0        SparseM_1.03    
[28] Hmisc_3.14-4     Formula_1.1-1    survival_2.37-7 
[31] sos_1.3-8        brew_1.0-6       lattice_0.20-29 

loaded via a namespace (and not attached):
 [1] caTools_1.16        cluster_1.15.2      colorspace_1.2-4   
 [4] descr_1.0.3         dichromat_2.0-0     digest_0.6.4       
 [7] fracdiff_1.4-2      gdata_2.13.3        gtable_0.1.2       
[10] gtools_3.4.0        hdrcde_3.1          KernSmooth_2.23-12 
[13] ks_1.9.1            labeling_0.2        latticeExtra_0.6-26
[16] Matrix_1.1-3        mgcv_1.7-29         misc3d_0.8-4       
[19] munsell_0.4.2       nlme_3.1-117        nnet_7.3-8         
[22] parallel_3.1.0      plyr_1.8.1          proto_0.3-10       
[25] quadprog_1.5-5      RColorBrewer_1.0-5  Rcpp_0.11.1        
[28] rgl_0.93.996        sandwich_2.3-0      scales_0.2.3       
[31] stringr_0.6.2       tools_3.1.0         tseries_0.10-32    
[34] xtable_1.7-3       


So that shows you how to provide some of the needed information (which neither of the postings to which you resonded had done.)  PLEASE read the Posting Guide.


>> Error in .C("spline_coef", method = as.integer(method), n = nx, x = x,
> : 
>        C symbol name "spline_coef" not in DLL for package "stats"
> 
> Cm.spline code from
> http://www-personal.buseco.monash.edu.au/~hyndman/Rlibrary/interpcode.R
> **********************************
> 
> I hope that helps, thank you :)! I've tried calling spline_coef from both 
> stats and base, but I get the same error for both packages. I've also tried 
> the solution suggested to these people, which was to remove any other R 
> versions on my laptop.
> 
> Best, 
> Lotte
> 
> 

David Winsemius
Alameda, CA, USA



More information about the R-help mailing list