# Need help forming a regexp ("not" operator?)
**URL:** https://boards.straightdope.com/t/need-help-forming-a-regexp-not-operator/579736
**Category:** Factual Questions
**Created:** [April 26, 2011, 3:20am UTC](https://boards.straightdope.com/t/need-help-forming-a-regexp-not-operator/579736 "2011-04-26T03:20:38Z")
**Posts on this page:** 8
**Page:** 1
### Author:  [@Jragon](https://boards.straightdope.com/u/Jragon)
#### Post date: [April 26, 2011, 3:20am UTC](https://boards.straightdope.com/t/need-help-forming-a-regexp-not-operator/579736/1 "2011-04-26T03:20:38Z")
I, unfortunately, never really learned regexps, and now I’m sort of in dire need of them. I could do this without them, but it would be a general PITA. I basically need to make a delimeter to match any characters NOT in a given set. \W gets close, but for the things I _don’t_ want to delimit I need to include a few special characters like apostrophes and hyphens, so something like --all the expressions NOT in “\w’-”–(that’s not all the characters necessary, but should give you an idea). You’d think it’d be easy to Google this, but it’s not. Maybe that’s because I don’t know how to use regexps…
---
### Author:  [@tellyworth](https://boards.straightdope.com/u/tellyworth)
#### Post date: [April 26, 2011, 3:23am UTC](https://boards.straightdope.com/t/need-help-forming-a-regexp-not-operator/579736/2 "2011-04-26T03:23:39Z")
[^abcd] means “match any character except a, b, c or d”.
So something like [^\w’-"–] should do what you want.
---
### Author:  [@yabob](https://boards.straightdope.com/u/yabob)
#### Post date: [April 26, 2011, 3:27am UTC](https://boards.straightdope.com/t/need-help-forming-a-regexp-not-operator/579736/3 "2011-04-26T03:27:39Z")
If a character class delimited with square braces beings with a caret, it means match something not in the class:
[^,;:]
Matches a single character which is not a comma semicolon of colon, for instance. - may also be used to indicate ranges:
[^A-G1-5]
Matches a single character not an uppercase A through G or 1 through 5.
---
### Author:  [@Jragon](https://boards.straightdope.com/u/Jragon)
#### Post date: [April 26, 2011, 3:33am UTC](https://boards.straightdope.com/t/need-help-forming-a-regexp-not-operator/579736/4 "2011-04-26T03:33:35Z")
Ah, I got confused because “^” seems to also be an anchor meaning “begins with.” What’s the difference between these usages? I assume it’s something like
^[A-G] means “starts with A-G”
while
[^A-G] mans “not A-G”
Is that right?
---
### Author:  [@yabob](https://boards.straightdope.com/u/yabob)
#### Post date: [April 26, 2011, 3:50am UTC](https://boards.straightdope.com/t/need-help-forming-a-regexp-not-operator/579736/5 "2011-04-26T03:50:25Z")
> [@Jragon](#):
>
> Ah, I got confused because “^” seems to also be an anchor meaning “begins with.” What’s the difference between these usages? I assume it’s something like
>
> \[1\] means “starts with A-G”
>
> while
>
> [^A-G] mans “not A-G”
>
> Is that right?
Yep.
* * *
1. A-G
---
### Author:  [@Jragon](https://boards.straightdope.com/u/Jragon)
#### Post date: [April 27, 2011, 5:48am UTC](https://boards.straightdope.com/t/need-help-forming-a-regexp-not-operator/579736/6 "2011-04-27T05:48:57Z")
So, I’m trying to make a delimiter that will separate paragraphs based on whether they have one or more blank lines separating them (in other words, 2 or more newline characters). I figured:
{2,}
Would work, when that failed, I tried
[
]{2,}
Which also didn’t work. Am I missing something? Is the scanner just being buggy?
---
### Author:  [@Jragon](https://boards.straightdope.com/u/Jragon)
#### Post date: [April 27, 2011, 6:02am UTC](https://boards.straightdope.com/t/need-help-forming-a-regexp-not-operator/579736/7 "2011-04-27T06:02:54Z")
Never mind, forgot Windows uses carriage returns. Stripping the file of carriage returns with a simple program (I only have to handle Unix-formatted files) worked perfectly.
---
### Author:  [@tellyworth](https://boards.straightdope.com/u/tellyworth)
#### Post date: [April 27, 2011, 6:06am UTC](https://boards.straightdope.com/t/need-help-forming-a-regexp-not-operator/579736/8 "2011-04-27T06:06:01Z")
Are you sure the input has just newlines (
) and not crlfs (
)?
ETA: never mind then.