Use a Regex quantifier in Awk

Question

I'm having trouble using quantifier in a Awk Regex.

awk -F"|" -v col="$COUNTER" '$col ~ /doi\/tm-[a-z0-9]/ { print $col } ' OFS=, tmp.csv

This works fine, but I want to specify to the [a-z0-9] class a limit of 8 chars.

So I did :

awk -F"|" -v col="$COUNTER" '$col ~ /doi\/tm-[a-z0-9]{8}/ { print $col } ' OFS=, tmp.csv

It doesn't work, and :

awk -F"|" -v col="$COUNTER" '$col ~ /doi\/tm-[a-z0-9]\{8\}/ { print $col } ' OFS=, tmp.csv

doesn't work neither.

That's on a Ubuntu 20.04 system.

Which awk implementation are you using? What Unix are you running on? Possibly related: unix.stackexchange.com/questions/561772/… — Kusalananda, Oct 16, 2020 at 14:30
I believe Ubuntu installs mawk by default which is not POSIX compliant and doesn't support regex interval operators (doesn't support POSIX character classes nor multibyte characters either). You can install gawk which should replace mawk with gawk for awk. Alternatively you could run busybox awk instead of awk. — Stéphane Chazelas, Oct 16, 2020 at 14:39

mgev · Accepted Answer · 2020-10-16 14:47:57Z

0

Indeed, changing default awk to gawk solved the problem.

gawk -v col="$COUNTER" 'BEGIN{ FS = "|" } $col ~ /doi\/tm-[a-z0-9]{8}/ { print $col } 1' OFS=, tmp.csv

This works fine.

answered Oct 16, 2020 at 14:47

mgev

91 bronze badge

Add a comment |

1 Answer 1