[R] error tracing

[Dieter Menne]

Steve Adams <steve_adams_sd <at> yahoo.com> writes:

 
> Hi, I am running some Cox modeling on large number of
> variables (thousands) using apply. For some reasons,
> some of the variables have problems in Cox regression,
> so the the run was stopped automatically. Is there a
> way to keep the running for all the variables to
> complete and let us, when done, to trace back which of
> the variables cause the problem?

Since each Cox model is a bit of work, you won't notice
the difference if you used a loop instead of apply. Within the for-loop, 
use try() to catch the bad guys.

Dieter