R help - Feb 2011 - How to flag those iterations which yield a warning?

Hi,

I am running a simulation study with the survival::coxph.  Some of the
simulations result in problematic fits due to flat partial likelihood. So, you
get the warning message:

Warning message:
In fitter(X, Y, strats, offset, init, control, weights = weights,  ... :
  Loglik converged before variable  2 ; beta may be infinite. 

How can I keep track of the simulations which yield any kind of warning message,
so that I can appropriately handle them in the post-processing of simulation
results?

I know that `try' can be used to capture errors, but I would like to capture
and detect warnings.

Any pointers would be appreciated.

Thanks,
Ravi.
____________________________________________________________________

Ravi Varadhan, Ph.D.
Assistant Professor,
Division of Geriatric Medicine and Gerontology
School of Medicine
Johns Hopkins University

Ph. (410) 502-2619
email: rvaradhan at jhmi.edu

May be I could do:

for (i in 1:nsim) {

last.warning <- NULL

# do model fitting

if(!is.null(last.warning)) # discard simulation result

}

I think this might work.  Any other ideas?

Ravi.

____________________________________________________________________

Ravi Varadhan, Ph.D.
Assistant Professor,
Division of Geriatric Medicine and Gerontology
School of Medicine
Johns Hopkins University

Ph. (410) 502-2619
email: rvaradhan at jhmi.edu


----- Original Message -----
From: Ravi Varadhan <rvaradhan at jhmi.edu>
Date: Friday, February 18, 2011 11:29 am
Subject: [R] How to flag those iterations which yield a warning?
To: r-help at r-project.org

> Hi,
>  
>  I am running a simulation study with the survival::coxph.  Some of 
> the simulations result in problematic fits due to flat partial 
> likelihood. So, you get the warning message:
>  
>  Warning message:
>  In fitter(X, Y, strats, offset, init, control, weights = weights,  
> ... :
>    Loglik converged before variable  2 ; beta may be infinite. 
>  
>  How can I keep track of the simulations which yield any kind of 
> warning message, so that I can appropriately handle them in the 
> post-processing of simulation results?
>  
>  I know that `try' can be used to capture errors, but I would like to 
> capture and detect warnings.
>  
>  Any pointers would be appreciated.
>  
>  Thanks,
>  Ravi.
>  ____________________________________________________________________
>  
>  Ravi Varadhan, Ph.D.
>  Assistant Professor,
>  Division of Geriatric Medicine and Gerontology
>  School of Medicine
>  Johns Hopkins University
>  
>  Ph. (410) 502-2619
>  email: rvaradhan at jhmi.edu
>  
>  ______________________________________________
>  R-help at r-project.org mailing list
>  
>  PLEASE do read the posting guide 
>  and provide commented, minimal, self-contained, reproducible code.

If you have a function that may produce warnings you
can use withCallingHandlers() to make a new function
that does what the original does but adds some
contextual information to the usual warning report.
E.g.,

addWarningReporting <- function(FUN) {
   force(FUN)
   function(args) {
      withCallingHandlers(
         FUN(args),
         warning=function(w){
             cat("WARNING: ")
             print(w)
             cat("    args=", deparse(args), "\n")
             invokeRestart("muffleWarning")}
      )
   }
}
> for(i in -2:2) log(i)
Warning messages:
1: In log(i) : NaNs produced
2: In log(i) : NaNs produced
> myLog <- addWarningReporting(log)
> for(i in -2:2) myLog(i)
WARNING: <simpleWarning in FUN(args): NaNs produced>
    args= -2L
WARNING: <simpleWarning in FUN(args): NaNs produced>
    args= -1L

Bill Dunlap
Spotfire, TIBCO Software
wdunlap tibco.com  
> -----Original Message-----
> From: r-help-bounces at r-project.org 
> [mailto:r-help-bounces at r-project.org] On Behalf Of Ravi Varadhan
> Sent: Friday, February 18, 2011 8:29 AM
> To: r-help at r-project.org
> Subject: [R] How to flag those iterations which yield a warning?
> 
> Hi,
> 
> I am running a simulation study with the survival::coxph.  
> Some of the simulations result in problematic fits due to 
> flat partial likelihood. So, you get the warning message:
> 
> Warning message:
> In fitter(X, Y, strats, offset, init, control, weights = 
> weights,  ... :
>   Loglik converged before variable  2 ; beta may be infinite. 
> 
> How can I keep track of the simulations which yield any kind 
> of warning message, so that I can appropriately handle them 
> in the post-processing of simulation results?
> 
> I know that `try' can be used to capture errors, but I would 
> like to capture and detect warnings.
> 
> Any pointers would be appreciated.
> 
> Thanks,
> Ravi.
> ____________________________________________________________________
> 
> Ravi Varadhan, Ph.D.
> Assistant Professor,
> Division of Geriatric Medicine and Gerontology
> School of Medicine
> Johns Hopkins University
> 
> Ph. (410) 502-2619
> email: rvaradhan at jhmi.edu
> 
> ______________________________________________
> R-help at r-project.org 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.
>