awk gensub got the wrong part of matches

Question

I want to extract \"yy a1 b1\" with the awk expression

echo \"xx yy [a1 b1] / zz [a2 b2]/\" | awk \'{p=gensub(/.*\\[([a-z0-9 ]+)\\].*/,\"\\\\1\",1); print $2,p}\         


        

Answer 1

The point is that the first .* as many chars as possible, thus making the regex engine match the last occurrence of the remaining subpatterns.

You may use the following solution:

/[^][]*\[([a-z0-9 ]+)\].*/

where [^][]* matches any 0 or more chars other than ] and [, thus, allowing the regex engine to stop right before the first [...]

See the online demo