What are some PRACTICAL scenarios for using the regex + symbol?

Question

So, I think I know what the RegEx + symbol does (looks for the preceding character 1 or more times). I think I get how to use it...

echo "This is a good sentence." | awk '/go+d/ {print $0}'

(finds good, gooood, gooooooood)

But, I'm having trouble coming up with a practical, objective reason for wanting to look for this kind of repetition. I need scenarios to give its purpose some context for explanation to others.

Thanks for helping and letting me lean on your knowledge and experiences.

I thought that too, but I couldn't think of a scenario to explain how that would look. — dlowrie290, Oct 10 '17 at 02:22
What about when you want to slurp up whitespace? At least one space, for example. `\s+`. Think about character classes rather than individual characters. — B Layer, Oct 10 '17 at 02:28
That is EXACTLY the kind of thing I'm looking for. A great example of why you would use this type of functionality. — dlowrie290, Oct 10 '17 at 02:33

B Layer · Accepted Answer · 2017-10-10T03:29:36.227

4

You're thinking about individual characters when you'll find that + is much more useful with character classes...

Whitespace: \s+
Non-numerical characters: [^0-9]+
A word: \w+
Non-empty string: .+
Etc.

Practical examples should be fairly easy to think of with these. Perhaps I want to find only lines that have indented text: ^\s+\S. Or I might want to find all non-empty lines: ^.+$. I can come up with more if you need them. ;)

edited Oct 10 '17 at 03:29

answered Oct 10 '17 at 02:42

B Layer

5,107
1
17
34

No, `[0-9]` is numerical. `[^0-9]` is NOT numerical. – B Layer Oct 10 '17 at 02:44
My bad, I read it as `^[`... – jasonwryan Oct 10 '17 at 02:45
No worries. I suspected that that's what you saw. :) – B Layer Oct 10 '17 at 02:45

score 3 · Answer 2 · answered Oct 10 '17 at 02:44

Here's a real-world instance of damage caused by using * instead of + (mostly because OP didn't realise that they would end up using regexes instead of wildcards): Why did this command delete every package?

Arguments to apt-get install and apt-get remove are extended regular expressions, not shell wildcards; wine* means win followed by any number of e, and since this can match any part of the package name, this means any package whose name contains win as a substring.

OP had run sudo apt-get remove wine*, ended up removing the majority of the installed packages in their system. wine+, on the other hand, would have removed packages with wine in it.

What are some PRACTICAL scenarios for using the regex + symbol?

2 Answers2