Regex variable length lookbehind into wildcard

Question

I have strings in the form of

[Identifier][Keyword][Number]Text

Where identifier is \w*, Keyword is one of Button, Link, or Text, Number is [0-9]* and text is the literal word Text. I am trying to match strings that don't fit this pattern.

Examples I would like matched:

MyIdentifier
MyIdentifierText
MyIdentifierButton
MyIdentifierButton1
Button
Button1
Text

Examples I don't want matched:

ButtonText
MyIdentifierButtonText
Button1Text
MyIdentifierButton1Text
TextText
Text1Text

I have come up with this:

^\w*(?:(?<!Button)|(?<!Link)|(?<!Text))[0-9]*(?<!Text)$

That is correct except it does not match the following:

Text
MyIdentifierText

I know the issue is with the first lookbehind, but I'm not sure how to fix it. I have read this and it didn't help, but that question also didn't have a wildcard before the lookbehind.

Just split a string with your matching pattern to get all non-matching parts. — Wiktor Stribiżew, Apr 11 '19 at 19:17
@WiktorStribiżew I'm using this to search for lines in files using Visual Studio so manipulating the strings is not an option. — Michael Brandon Morris, Apr 11 '19 at 19:22
Visual Studio does not use PCRE, it uses .NET regex flavor. Do you have these strings on separate lines? If yes, try `^(?!\w*(?:Button|Link|Text)[0-9]*Text$).*` — Wiktor Stribiżew, Apr 11 '19 at 19:24
@WiktorStribiżew My bad on the tag, apparently there isn't one for .NET regex (or at least I couldn't find it). I'm testing using a PCRE tool and the sample data is on separate lines for now. The actual strings are not necessarily on separate lines but are contained in HTML data attributes (data-myattribute="{string}"). I felt including that detail would clutter the question. — Michael Brandon Morris, Apr 11 '19 at 19:31
Using a PCRE testing tool won't help if the ultimate goal is to do this in VS. — Barmar, Apr 11 '19 at 19:54
Please provide exact sample data and exact expected output. I understand you are using a search and replace dialog. — Wiktor Stribiżew, Apr 12 '19 at 16:25
If you want to remove all text not matching your pattern, it is much simpler in Notepad++. — Wiktor Stribiżew, Apr 12 '19 at 17:43
Try `(\w*(?:Button|Link|Text)[0-9]*Text)|(?:(?!\w*(?:Button|Link|Text)[0-9]*Text)[\s\S\r])+` and replace with `$1\n`, see [this demo](https://regex101.com/r/QnFmqX/1). — Wiktor Stribiżew, Apr 12 '19 at 17:45
Please leave feedbak if you need any more hints on it. It is not quite clear what you are doing. — Wiktor Stribiżew, Apr 27 '19 at 07:14

Regex variable length lookbehind into wildcard

0 Answers0