pdftk extract random pages from script variable

I have some pdf files about 2000 pages. They are randomly generated. I need to extract some pages that contain some specific patterns, that changes its page number for every pdf. With some steps I can get the page numbers and I store some info into a csv file like that:

PatternA ; 1 3 5 7 PatternB ; 1 8 10 22

I have been trying to do a loop to get and process each line from this csv into the cat option from pdftk command, but it aways return error:

$IFS=$(printf '\n\t') for line in cat job.csv do pattern=echo $line ¦ cut -d ';' -f 1 pages=echo $line ¦ cut -d ';' - f 2 pdftk input.pdf cat $pages output $pattern done

When echoing pattern and pages variables, everything are ok. But the pdftk command returns error if I try to get pages from $pages variable.

What I am doing wrong?

Thanks!