I want a regex matching a specific word that is not surrounded by any alphanumeric character. My thought was to include a negation before and after:

[^a-zA-Z\d]myspecificword[^a-zA-Z\d]

So it would match:

myspecificword
_myspecificword_
-myspecificword

And not match:

notmyspecificword
myspecificword123

But this simple regex won't match the word by itself unless it is preceeded by a whitespace:

myspecificword   // no match
 myspecificword  // match

Using the flags "gmi" and testing with JavaScript. What am I doing wrong? Shouldn't it be as simple as that?

I want a regex matching a specific word that is not surrounded by any alphanumeric character. My thought was to include a negation before and after:

[^a-zA-Z\d]myspecificword[^a-zA-Z\d]

So it would match:

myspecificword
_myspecificword_
-myspecificword

And not match:

notmyspecificword
myspecificword123

But this simple regex won't match the word by itself unless it is preceeded by a whitespace:

myspecificword   // no match
 myspecificword  // match

Using the flags "gmi" and testing with JavaScript. What am I doing wrong? Shouldn't it be as simple as that?

https://regex101./r/BCkbVQ/3

• javascript
• regex

• 1 Regex word boundary is \b – user47589 Commented Feb 28, 2020 at 16:05
• @Amy That won't work because underscore is considered a word character. – Barmar Commented Feb 28, 2020 at 16:08
• Use negative lookarounds. regular-expressions.info/lookaround.html – Barmar Commented Feb 28, 2020 at 16:08
• 3 It doesn't work because [^a-zA-Z\d] needs to match an actual character. There's no character at the beginning or end. – Barmar Commented Feb 28, 2020 at 16:09
• /(?<![a-z\d])myspecificword(?![a-z\d])/ig should work but keep in mind lookbehind is not supported in older browsers. – anubhava Commented Feb 28, 2020 at 16:16

5 Answers 5

Try using:

(?<![^\s_-])myspecificword(?![^\s_-])

This says to match myspecificword when it surrounded, on both sides, by either the start/end of the input, whitespace, underscore, or dash.

Demo

It is not whitespace that is required but any symbol that is matches [^a-zA-Z\d].

You should use: (Demo)

(?:^|[^a-zA-Z\d])myspecificword(?:[^a-zA-Z\d]|$)

The main benefit is support across all Regexp parsers.

If you truly mean "not surrounded by alphanumerics other than _ (and in your attempted regex you seem to be willing to match anything that isn't a letter or digit), then any of the following should be acceptable:

'myspecificword'
'_myspecificword_'
' myspecificword '
'-myspecificword-'
'(myspecificword)'

And the regex should be:

(?<![^_\W])myspecificword(?![^_\W])

let tests = ['myspecificword',
             '_myspecificword_',
             ' myspecificword ', 
             '-myspecificword-', 
             '(myspecificword)', 
             'amyspecificword',
             '1myspecificword'
             ];
let regex = /(?<![^_\W])myspecificword(?![^_\W])/;
for (let test of tests) {
    console.log(regex.test(test));
}

The "accepted" answer will not match (myspecificword), for example.

The title of this question is

Regex for word not surrounded by alphanumeric characters

The other answers have all addressed a different question (which may well be the one intended):

Regex for word neither preceded nor followed by alphanumeric characters

I will refer to these statements as #1 and #2 respectively.

If the specified word were 'cat' and the string were '9cat', 'cat' is not surrounded by alphanumeric characters in the string, so there is a match with #1, but not with #2.

For #1, one could use the regex:

/cat(?!\p{Alpha}|(?<!\p{Alnum})cat/

("match 'cat' not followed by a Unicode alphanumeric character or 'cat' not preceded by a Unicode alphanumeric character"), though it's easier to test for the negation:

/(?<=\p{Alpha}cat(?<=\p{Alnum})/

The test passes if the string does not match this regex.

With interpretation #2, the regex is:

/(?<!\p{Alpha}cat(?!\p{Alnum})/

I think this will work:

/[^a-z0-9]?myspesificword[^a-z0-9]?/i