increment var in gnu parallel

Question

How can I use an increment variable within parallel? Note the $int variable in the output filename (prefix). I realize order can change, and that is fine, but what's useful is to have integers prefixing the output for downstream work (the length of ${array[@]} changes).

I'm trying to convert this for-loop into a nice parallel cmd to run on a cluster. One feature I really like is to have this process quit if any of the jobs have a non-zero exit status described here: https://www.gnu.org/software/parallel/parallel_tutorial.html#Termination

From:

var='infile.txt'
int=1
for file in ${array[@]}; do
  script.rb -i $var -o $int_$file
  ((int+=1))
done

To:

parallel --halt soon,fail=1 --jobs $NSLOTS script.sh -i $var -o int_{} ::: ${array[@]}

I was thinking something like a {1..array_len} construct might work to act as a counter but this doesn't: parallel --halt soon,fail=1 --jobs $NSLOTS script.sh -i $var -o {2}_{1} :::: ${array[@]} {1..${#array[@]}}

Answer 1

I think you want this:

parallel --halt soon,fail=1 script.rb -i $var -o {#}_{} ::: ${array[@]}

where {#} is the job sequence number.

---

Here is a simple example:

parallel -k echo "param="{} ",sequence="{#} ::: 2 4 6 8

param=2 ,sequence=1
param=4 ,sequence=2
param=6 ,sequence=3
param=8 ,sequence=4