***caught segfault*** caused by Error in names(fit$coef) <- cname : 'names' attribute [4] must be the same length as the vector [2] about jm HOT 5 CLOSED

Furthermore, when the segfault error will be thrown might depend on ulimit size.
After setting ulimit -Ss unlimited, previously, it is 8192, the segfault error throws at a later iteration i = 87.

from jm.

This is a pathogenic situation for which you would not expect to fit a joint model. The problem is caused by survival::survreg which returns NAs as estimated coefficients for which it fails to set names.

from jm.

Thanks for your hint.

Could you explain why there is a segfault error? In practice, I want to conduct joint model for multiple variables, just like the for loop I used in the above toy example. I have wrapped jointModel() with try{}, so I expected that it is OK if it failed, but as you can see, the loop completely broken down after several iterations, and throws a segfault error.

from jm.

The segfault is caused in survival:::survreg.fit that is called by the JM::jointModel(). Hence, you should ask the maintainer of the survival package. But to stress again, this is a pathological situation.

from jm.

Thanks for your suggestion! I understand this is a quite extreme case, but without any prior knowledge, I want the program can be finished smoothly, regardless of bad fitting or good fitting. However currently, the segfault error breaks down the program, even I have used try.

from jm.