Hi Lachlan,
On Thu, Oct 26, 2017 at 10:10 AM, Lachlan Musicman <data...@gmail.com> wrote: > for fasta_file in `ls /path/to/fasta_files`; do sbatch myscript.sbatch > $fasta_file; done > > They would like their output and error files to be like: > > --output=$fasta_file.out > --error=$fasta_file.err I analyze a lot of FASTA-like files and I don't see a very big problem. Perhaps your users and I think quite differently? With the above loop, $fasta_file would be the first argument to myscript.sbatch. I would take that and form stdout and stderr filenames. Whatever program they are running in that script (i.e., blast or some other aligner) can then write to stdout and stderr with these filenames. Yes...true. It takes a bit more work but then again, your users will just need to do this programming once. My scripts actually have this at the top: #SBATCH --output="%j-%N.out" which gives me the job number and computational node (as I've piped stdout and stderr of the programs to another file, this file is more or less empty). IMHO (again, maybe I think differently from your users), with this information, I can figure out information about the job and the node that it was run. I think this information is important in case reproducibility is important. Sorry, not quite what you're looking for but I hope it helps your users a bit. Ray
