R help - May 2005 - Error handling

Hi,

I have written an R-Procedure that automatically does a ML-Fit for some
datasets to a few
selected distributions. After any optimization the function writes the
estimated parameters into a variable.
My problem is now that sometimes the optimization failes and my program
stops with an error message.

What I want after a failed optimization is that there is written an error
msg
in the variable instead of stopping the program.

For the ML-Fit I use the "fitdistr" algorithm.

How could I do that ?

Thank you very very much!

Carsten

On 5/23/05, Carsten Steinhoff <carsten.steinhoff at gmx.de>
wrote:
> Hi,
> 
> I have written an R-Procedure that automatically does a ML-Fit for some
> datasets to a few
> selected distributions. After any optimization the function writes the
> estimated parameters into a variable.
> My problem is now that sometimes the optimization failes and my program
> stops with an error message.
> 
> What I want after a failed optimization is that there is written an error
> msg
> in the variable instead of stopping the program.
> 
> For the ML-Fit I use the "fitdistr" algorithm.
> 
See

?try
?tryCatch

For example, using try:

result <- try(myfun(...whatever...))
if(inherits(result, "try-error")) ...process the error...