[R] Catch errors

[Gang Chen]

Seth and Stephen,

Thanks a lot for the explanation and help! I really appreciate it.

Gang


On Aug 6, 2007, at 6:41 PM, Stephen Tucker wrote:
> That's because tag <- 1 is evaluated in a local environment (of the  
> function)
> - once function(err) exits, the information is lost. Try R's <<-  
> operator:
>
> tag <- 0
> tryCatch(fit.lme <- lme(Beta ~ Trust*Sex*Freq, random = ~1|Subj,
>          Model), error=function(err) tag <<- 1)


On Aug 6, 2007, at 6:55 PM, Seth Falcon wrote:

> Gang Chen <gangchen at mail.nih.gov> writes:
>
>> I wanted something like this:
>>
>>> tag <- 0;
>>> tryCatch(fit.lme <- lme(Beta ~ Trust*Sex*Freq, random = ~1|Subj,
>> Model), error=function(err) tag <- 1);
>>
>> but it seems not working because 'tag' does not get value of 1 when
>> error occurs. How can I make it work?
>
> You can use '<<-' to assign to the parent frame like this:
>
>         tag <- 0
>         tryCatch({
>             print("inside tryCatch")
>             print(paste("tag", tag))
>             stop("forcing an error")
>         }, error=function(e) {
>             print("caught an error")
>             tag <<- 1
>         })
>
> But you can also use the return value of tryCatch as a return code if
> that is what "tag" is.  Like this:
>
>         fail <- TRUE
>         tag <- tryCatch({
>             print("inside tryCatch")
>             if (fail)
>               stop("forcing an error")
>             0
>         }, error=function(e) {
>             print("caught an error")
>             1
>         })
>         tag
>
> + seth
>
> -- 
> Seth Falcon | Computational Biology | Fred Hutchinson Cancer  
> Research Center
> BioC: http://bioconductor.org/
> Blog: http://userprimary.net/user/