R help - Mar 2008 - Conditionals for Errors or error flags

Greetings,

I have been working on a script that conducts repeated statistics and  
plots to my data.  In this case it is sub-setting the dataframe by  
month.
The intent is to develop a custom analysis and plotting that I can run  
on a large number of data sets.

Unfortunately, a small portion of my subsets (~1%) cause an error with  
one of the wrapped subroutines that results in the whole routine  
aborting.  It would be incredibly difficult to find the cause of this  
error post-run, and the particular routine does not have a way of  
treating the error (it isn't as simple as NA, etc.)

What I am wondering is if there is a corollary to the conditional  
is.na() such as is.error()? Or is there a way to get some other output  
from an error than an abort?

That I can use to toggle between a process that causes an error and  
one that doesn't without aborting the whole shebang.

Any ideas are appreciated.


I apologize that I could not think up an example so I included a  
psuedo-code below.

do.something<-function(x){
	run other scripts with data, possibly causing an error
}

do.something.else<-function(x){
	Something that doesn't cause an error
}

run.prog<-function(x){

	if(IS.ERROR(do.something)){do.something.else}else{do.something}  #or  
flag error
}

run.prog(data)
#runs everything if there is an error it does not abort with error but  
rather does something else that doesn't error and continues.
	
Thank you very much,

Jon
R 2.6.2
MAC OS 10.5

Jon Loehrke
Graduate Research Assistant
Department of Fisheries Oceanography
School for Marine Science and Technology
University of Massachusetts
200 Mill Road, Suite 325
Fairhaven, MA 02719
jloehrke at umassd.edu
508-758-6393

Dear Jon,

You probably want to take a look at "try" and "tryCatch". 
Either of
them will let you do what you want.

Best,

R.

On Sat, Mar 15, 2008 at 7:00 PM, Jon Loehrke <jloehrke at umassd.edu>
wrote:
> Greetings,
>
>  I have been working on a script that conducts repeated statistics and
>  plots to my data.  In this case it is sub-setting the dataframe by
>  month.
>  The intent is to develop a custom analysis and plotting that I can run
>  on a large number of data sets.
>
>  Unfortunately, a small portion of my subsets (~1%) cause an error with
>  one of the wrapped subroutines that results in the whole routine
>  aborting.  It would be incredibly difficult to find the cause of this
>  error post-run, and the particular routine does not have a way of
>  treating the error (it isn't as simple as NA, etc.)
>
>  What I am wondering is if there is a corollary to the conditional
>  is.na() such as is.error()? Or is there a way to get some other output
>  from an error than an abort?
>
>  That I can use to toggle between a process that causes an error and
>  one that doesn't without aborting the whole shebang.
>
>  Any ideas are appreciated.
>
>
>  I apologize that I could not think up an example so I included a
>  psuedo-code below.
>
>  do.something<-function(x){
>         run other scripts with data, possibly causing an error
>  }
>
>  do.something.else<-function(x){
>         Something that doesn't cause an error
>  }
>
>  run.prog<-function(x){
>
>         if(IS.ERROR(do.something)){do.something.else}else{do.something} 
#or
>  flag error
>  }
>
>  run.prog(data)
>  #runs everything if there is an error it does not abort with error but
>  rather does something else that doesn't error and continues.
>
>  Thank you very much,
>
>  Jon
>  R 2.6.2
>  MAC OS 10.5
>
>  Jon Loehrke
>  Graduate Research Assistant
>  Department of Fisheries Oceanography
>  School for Marine Science and Technology
>  University of Massachusetts
>  200 Mill Road, Suite 325
>  Fairhaven, MA 02719
>  jloehrke at umassd.edu
>  508-758-6393
>
>  ______________________________________________
>  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.
>


-- 
Ramon Diaz-Uriarte
Statistical Computing Team
Structural Biology and Biocomputing Programme
Spanish National Cancer Centre (CNIO)
http://ligarto.org/rdiaz