Lazy regex operator doesn't work in bash

Question

echo "$(expr "title: Purple Haze       artist: Jimi Hendrix" : 'title:\s*\(.*\?\)\s*artist.*' )"

prints

Purple Haze

With the trailing whitespace, even though I am using the ? lazy operator.

I've tested this on https://regex101.com/ and it works as expected, what's different about bash?

@Enissay No, these are BRE, not ERE: `?` matches a question mark, `\?` is the zero-or-one operator. — Gilles 'SO- stop being evil', Jan 04 '15 at 15:24
@JakubM. Sorry I should have added, I am on Debian and `expr` on linux is a bit more powerful — texasflood, Jan 04 '15 at 15:30

score 6 · Accepted Answer · answered Jan 04 '15 at 15:23

6

You aren't using bash's regexp matching, you're using expr. expr does not have a “? lazy operator”, it only implements basic regular expressions (with a few extensions in the Linux version, such as \s for whitespace, but that doesn't include Perl-like lazy operators). (Neither does bash, for that matter.)

If you don't want .* to include trailing space, specify that it must end with a character that isn't a space:

'title:\s*\(.*\S\)\s*artist.*'

answered Jan 04 '15 at 15:23

Gilles 'SO- stop being evil'

104,111
38
209
254

Perhaps also mention that `expr` should generally be avoided. If you are using Bash, it already has built-in replacements for everything `expr` does, which are often also more readable, efficient, versatile, and robust. – tripleee May 03 '23 at 08:29

score 2 · Answer 2 · edited May 23 '17 at 12:28

As Gilles points out, you're not using bash regular expressions. To do so, you could use the regex match operator =~ like this:

re='title:[[:space:]]*(.*[^[:space:]])[[:space:]]*artist.*'
details='title: Purple Haze       artist: Jimi Hendrix'
[[ $details =~ $re ]] && echo "${BASH_REMATCH[1]}"

Rather than using a lazy match, this uses a non-space character at the end of the capture group, so the trailing space is removed. The first capture group is stored in ${BASH_REMATCH[1]}.

At the expense of cross-platform portability, it is also possible to use the shorthand \s and \S instead of [[:space:]] and [^[:space:]]:

re='title:\s*(.*\S)\s*artist.*'

Lazy regex operator doesn't work in bash

2 Answers2

Linked

Related