How to keep the parallel for loop running even when one of the job fails

Question: I run multiple jobs (each job is independent of each other) in parallel mode. However, if one job fails, then entire parfor loop is shut down.

More information : To better explain it, here's an example from my code:

*parfor* caseNumber  = 1:100

    % _For every case number, perform the following task_

    Step 1) Load input file for this case. Note: Each case has a separate file
    Step 2) Do some post-processing on it

*end*

However, there are cases in which input is such that we can have some error when we do some post-processing. In short, the post-processing step (i.e. Step 2) fails. At this stage, the entire parfor loop is shut and the code terminates.

But, I want the loop to continue since there are multiple cases for which post processing needs to be carried out. What is the best way to do it

[Edric Ellis]

What do you mean by "fail"? MATLAB error, worker crash, ...? If it's just an error, can you use try/catch around step 2?

[Sam T]

Good suggestion, Edric. It works.