Question

I have a regular expression for use with awk to find any of the specified words in a line of a file. It looks like this awk "/word1/||/word2/||/word3/" filename. As an alternative, I have been trying to specify the words like this WORDS="word1 word2 word3" and then use bash string substitution to form the regular expression to pass to awk.

I’ve tried numerous ways of doing this to no avail. awk simply dumps the contents of the entire file or spits out some complaint about the regex form.

Here’s what I have:

#!/bin/bash

FILE="myfile"
WORDS="word1 word2 word3"

# use BASH string substitution to obtain the regex which should look like this:
# "/word1/||/word2/||/word3/"

REGEX=\"/${WORDS// //||/}/\"

awk ${REGEX} $FILE

I’m fairly sure it has to do with quoting and I’ve tried various methods using echo and back ticks and can get it look right (when echoed) but when actually trying to use it, it fails.

Answer 1

Try to replace:

REGEX=\"/${WORDS// //||/}/\"

with:

REGEX="/${WORDS// //||/}/"

Note that there is no need to escape double quotes since they are not really part of the regular expression.