How to handle duplicate records

[biermanr]

There could be rare situations where it's possible for two different SLURM jobs can have the same jobid such as if the SLURM scheduler "wraps around" and starts over at jobid 1. This could lead to a situation where two separate jobs in the slurmise database might have the same jobid. I don't know if this would cause an issue, and even if it did this even would be rare so this is not a pressing concern.

[iparask]

This can happen with any path that a record may have in the HDF5 file. A first approach would be to set a flag to find a way to allow duplicate paths for records.

[iparask]

We do currently have a flag to ignore_existing_job that ignores the record when the job exists and the job data is exactly the same.

This is more where this is a duplication we did not account for.