[R] validate (package Design): error message "subscript out of bounds"

Dirk Eddelbuettel edd at debian.org
Tue Aug 28 15:22:37 CEST 2007


On 28 August 2007 at 08:04, Frank E Harrell Jr wrote:
| Wentzel-Larsen, Tore wrote:
| > Thanks,
| > I use Design version 2.1-1 (and as provided, R 2.5.1 on Windows XP).
| 
| Sorry I missed the latter.
| 
| > The redundancies in object names were for my own convience, as part of 
| > a larger command file, and the validation of this univariate model was
| > only included to ease comparison with the main multivariate model. I have 
| 
| A minor point: adding more variables to the right hand side makes the 
| model a multivariable model.  It is still univariate as it has only one 
| dependent variable.
| 
| > tried to access Design version 2.0_12, but only managed to access version
| > 2.1_1 of Design in my Windows implementation of R2.5.1 (the choice of 
| > operative system is made by my institution and I am only entitled to use 
| > Windows).
| 
| My mistake again.  I forgot that the Debian repositories for CRAN are 
| behind. 

In which way?  I se that ...

edd at ron:~$ wajig policy r-cran-design
r-cran-design:
  Installed: (none)
  Candidate: 2.1-1-1
  Version table:
     2.1-1-1 0
        500 http://ron testing/main Packages
         -1 http://ron unstable/main Packages
edd at ron:~$ wajig policy r-cran-hmisc
r-cran-hmisc:
  Installed: 3.4-2-2
  Candidate: 3.4-2-2
  Version table:
 *** 3.4-2-2 0
        500 http://ron testing/main Packages
         -1 http://ron unstable/main Packages
        100 /var/lib/dpkg/status
edd at ron:~$ 

... both Hmisc and Design are at current version in both testing and
unstable, and these are the versions on CRAN.  What am I missing?

Thanks, Dirk 

| 
| Frank
| 
| > Best, Tore
| > 
| > -----Opprinnelig melding-----
| > Fra: Frank E Harrell Jr [mailto:f.harrell at vanderbilt.edu] 
| > Sendt: 28. august 2007 03:17
| > Til: Wentzel-Larsen, Tore
| > Kopi: r-help at stat.math.ethz.ch
| > Emne: Re: [R] validate (package Design): error message "subscript out of bounds"
| > 
| > Wentzel-Larsen, Tore wrote:
| >> Dear R users 
| >>
| >> I use Windows XP, R2.5.1 (I have read the posting guide, I have 
| >> contacted the package maintainer first, it is not homework).
| >>
| >> In a research project on renal cell carcinoma we want to compute 
| >> Harrell's c index, with optimism correction, for a multivariate 
| >> Cox regression and also for some univariate Cox models.
| >> For some of these univariate models I have encountered an error
| >> message (and no result produced) from the function validate i 
| >> Frank Harrell's Design package:
| >>
| >> Error in Xb(x[, xcol, drop = FALSE], coef, non.slopes, non.slopes.in.x,  : 
| >>         subscript out of bounds
| >>
| >> The following is an artificial example wherein I have been able to 
| >> reproduce this error message (actual data has been changed to preserve
| >> confidentiality):
| > 
| > I could not reproduce the error on R 2.5.1 on linux using version 2.0-12 
| > of Design (you did not provide this information).
| > 
| > Your code involved a good deal of extra typing.  Here is a streamlined 
| > version:
| > 
| > bc <- data.frame(time1 = c(9,24,28,43,58,62,66,107,116,118,123,
| > 	127,129,131,137,138,139,140,148,169,176,179,188,196,210,218,
| > ....
| > bc
| > 
| > library(Design)
| > 
| > dd <- with(bc, datadist(bc1, age, adjto.cat='first'))
| > options(datadist = 'dd')
| > 
| > f <- cph(Surv(time1,status1) ~ bc1,
| >           data = bc, x=TRUE, y=TRUE, surv=TRUE)
| > anova(f)
| > f
| > summary(f)
| > 
| > val <- validate(f, B=200, dxy=TRUE)
| > 
| > I don't get much value of putting the type of an object as part of the 
| > object's name, as information within objects defines the object type/class.
| > 
| > There is little reason to validate a one degree of freedom model.
| > 
| > Frank
| > 
| >> library(Design)
| >>
| >> # an example data frame:
| >> frame.bc <- data.frame(time1 = c(9,24,28,43,58,62,66,107,116,118,123,
| >> 	127,129,131,137,138,139,140,148,169,176,179,188,196,210,218,
| >> 	1,1,1,2,2,3,4,8,23,32,33,34,43,44,48,51,52,54,59,59,60,60,62,
| >> 	65,65,68,70,72,73,74,81,84,88,98,99,106,107,115,115,117,119,
| >> 	120,122,122,122,122,126,128,130,135,136,136,138,149,151,154,
| >> 	157,159,161,164,164,164,166,172,172,176,179,180,183,183,184,
| >> 	187,190,197,201,201,203,203,203,209,210,214,219,227,233,4,18,
| >> 	49,113,147,1,1,2,2,2,2,2,3,4,6,6,6,6,6,6,6,6,9,9,9,9,9,10,10,
| >> 	10,11,12,12,12,13,14,14,17,18,18,19,19,20,20,21,21,21,21,22,23,
| >> 	23,24,28,28,29,29,32,34,35,38,38,48,48,52,52,54,54,56,64,67,67,
| >> 	69,70,70,72,84,88,90,114,115,140,142,154,171,195),
| >> 	status1 = c(0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
| >> 	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
| >> 	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
| >> 	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
| >> 	1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
| >> 	1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
| >> 	1,1,1,1,1),
| >> 	bc1 = factor(c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
| >> 	2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,
| >> 	2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,
| >> 	2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,
| >> 	2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,
| >> 	2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2),
| >> 	labels=c('bc.1','bc.2')),
| >> 	age = c(58,68,23,20,50,43,41,69,20,48,19,27,39,20,65,49,70,59,31,43,25,
| >> 	61,60,45,34,59,32,58,30,62,26,44,52,29,40,57,33,18,50,50,55,51,38,34,
| >> 	69,56,67,38,66,21,48,39,62,62,29,68,66,19,60,39,55,42,24,29,56,61,40,
| >> 	52,19,40,33,67,66,51,48,63,60,58,68,60,53,20,45,62,37,38,61,63,43,67,
| >> 	49,39,43,67,49,69,32,37,32,63,33,47,66,39,23,57,26,61,20,49,69,30,40,
| >> 	29,38,66,60,69,69,44,65,25,41,53,18,55,45,59,49,27,51,29,67,26,24,26,
| >> 	47,23,50,27,35,45,32,26,45,45,63,39,39,22,38,27,31,27,49,65,66,49,39,
| >> 	21,51,49,55,63,19,26,50,21,24,34,65,33,55,33,36,53,48,25,54,58,60,34,
| >> 	47,23,34,60,39,34,22,30,41,55,64,48,34,54))
| >> frame.bc
| >>
| >> # preparing for a simple univariate Cox regression:
| >> dd.bc <- datadist(frame.bc[, c('bc1','age')], adjto.cat='first')
| >> options(datadist = 'dd.bc')
| >>
| >> # a univariate Cox regression:
| >> cph.bc <- cph(formula = Surv(time1,status1)~bc1,
| >> 	data = frame.bc, x=TRUE, y=TRUE, surv=TRUE)
| >> anova(cph.bc)
| >> cph.bc
| >> summary(cph.bc)
| >>
| >> # the validate command for the Cox model:
| >> val.cph.bc <- validate(cph.bc, B=200, dxy=TRUE , pr=TRUE)
| >>
| >> ----------------------
| >> Output from the validate command:
| >>
| >>        training       test
| >> Dxy   -0.124360 -0.1423409
| >> R2     1.000000  1.0000000
| >> Slope  1.000000  0.7919584
| >> D      0.016791  0.0147536
| >> U     -0.002395  0.0006448
| >> Q      0.019186  0.0141088
| >>        training       test
| >> Dxy   -0.191875 -0.1423409
| >> R2     1.000000  1.0000000
| >> Slope  1.000000  0.8936724
| >> D      0.022397  0.0147536
| >> U     -0.002339  0.0001367
| >> Q      0.024736  0.0146169
| >>        training       test
| >> Dxy   -0.199514 -0.1423409
| >> R2     1.000000  1.0000000
| >> Slope  1.000000  0.8075246
| >> D      0.025717  0.0147536
| >> U     -0.002447  0.0005348
| >> Q      0.028163  0.0142188
| >> Error in Xb(x[, xcol, drop = FALSE], coef, non.slopes, non.slopes.in.x,  : 
| >>         subscript out of bounds
| >>
| >>
| >> Any help/suggestions will be highly appreciated.
| >>
| >>
| >> Sincerely,
| >> Tore Wentzel-Larsen
| >> statistician
| >> Centre for Clinical research
| >> Armauer Hansen house 
| >> Haukeland University Hospital
| >> N-5021 Bergen
| >> tlf   +47 55 97 55 39 (a)
| >> faks  +47 55 97 60 88 (a)
| >> email tore.wentzel-larsen at helse-bergen.no
| >>
| >> ______________________________________________
| >> R-help at stat.math.ethz.ch mailing list
| >> https://stat.ethz.ch/mailman/listinfo/r-help
| >> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
| >> and provide commented, minimal, self-contained, reproducible code.
| >>
| > 
| > 
| 
| 
| -- 
| Frank E Harrell Jr   Professor and Chair           School of Medicine
|                       Department of Biostatistics   Vanderbilt University
| 
| ______________________________________________
| R-help at stat.math.ethz.ch mailing list
| https://stat.ethz.ch/mailman/listinfo/r-help
| PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
| and provide commented, minimal, self-contained, reproducible code.

-- 
Three out of two people have difficulties with fractions.



More information about the R-help mailing list