Return to Answer

Consider this:

cat ../benchmark/*/labels.txt |
while IFS=$'\t' read P1 P2 P3 ; do
  [[ $P3 == 76 ]] && echo $P1 $P2
done |
sort -u |
parallel ./build.oct

• you save awk subprocess with readline built-in parser (see comments below)
• you avoid dupes with sort -u
• you leverage resource usage with parallel (or xargs -l1)

Other approach of interest, piloted by awk:

awk -F'\t' '$3==76 && !seen[$1,$2]++ {
  print $1 FS $2 | "parallel"
}' ../benchmark/*/labels.txt

• reuse input field separator FS instead of literal
• dupes are discarded using an array of counters
• you learn piping to awk subprocess

Consider this:

cat ../benchmark/*/labels.txt |
while IFS=$'\t' read P1 P2 P3 ; do
  [[ $P3 == 76 ]] && echo $P1 $P2
done |
sort -u |
parallel ./build.oct

• you save awk subprocess with readline built-in parser (see comments below)
• you avoid dupes with sort -u
• you leverage resource usage with parallel (or xargs -l1)

Other approach of interest, piloted by awk:

awk -F'\t' '$3==76 && !seen[$1,$2]++ {
  print $1 FS $2 | "parallel ./build.oct"
}' ../benchmark/*/labels.txt

• reuse input field separator FS instead of literal
• dupes are discarded using an array of counters
• you learn piping to awk subprocess

Consider this:

cat ../benchmark/*/labels.txt |
while IFS=$'\t' read P1 P2 P3 ; do
  [[ $P3 == 76 ]] && echo $P1 $P2
done |
sort -u |
parallel ./build.oct

• you save awk subprocess with readline built-in parser
• you avoid doubles with sort -u
• you leverage resource usage with parallel (or xargs -l1)

Other approach of interest, piloted by awk:

awk -F'\t' '$3==76{
  print $1 FS $2 | "sort -u | parallel"
}' ../benchmark/*/labels.txt

• you learn piping to awk subprocess

Consider this:

cat ../benchmark/*/labels.txt |
while IFS=$'\t' read P1 P2 P3 ; do
  [[ $P3 == 76 ]] && echo $P1 $P2
done |
sort -u |
parallel ./build.oct

• you save awk subprocess with readline built-in parser (see comments below)
• you avoid dupes with sort -u
• you leverage resource usage with parallel (or xargs -l1)

Other approach of interest, piloted by awk:

awk -F'\t' '$3==76 && !seen[$1,$2]++ {
  print $1 FS $2 | "parallel"
}' ../benchmark/*/labels.txt

• reuse input field separator FS instead of literal
• dupes are discarded using an array of counters
• you learn piping to awk subprocess

Consider this:

cat ../benchmark/*/labels.txt |
while IFS='\t' read P1 P2 P3 ; do
  [[ $P3 == 76 ]] && echo $P1 $P2
done |
sort -u |
parallel ./build.oct

• you save awk subprocess with readline built-in parser
• you avoid doubles with sort -u
• you leverage resource usage with parallel (or xargs -l1)

Other approach of interest, piloted by awk:

awk -F'\t' '$3==76{
  print $1 FS $2 | "sort -u | parallel"
}' ../benchmark/*/labels.txt

• you learn piping to awk subprocess

Consider this:

cat ../benchmark/*/labels.txt |
while IFS=$'\t' read P1 P2 P3 ; do
  [[ $P3 == 76 ]] && echo $P1 $P2
done |
sort -u |
parallel ./build.oct

• you save awk subprocess with readline built-in parser
• you avoid doubles with sort -u
• you leverage resource usage with parallel (or xargs -l1)

Other approach of interest, piloted by awk:

awk -F'\t' '$3==76{
  print $1 FS $2 | "sort -u | parallel"
}' ../benchmark/*/labels.txt

• you learn piping to awk subprocess