How to properly escape characters in regexp

Question

I want to do a string search inside a string. Simply saying MySTR.search(Needle).

The problem occurs when this needle string contains special regex characters like *,+ and so on. It fails with error invalid quantifier.

I have browsed the web and found out that string can be escaped with \Q some string \E.

However, this does not always produce the desired behavior. For example:

var sNeedle = '*Stars!*';
var sMySTR = 'The contents of this string have no importance';
sMySTR.search('\Q' + sNeedle + '\E');

Result is -1. OK.

var sNeedle = '**Stars!**';
var sMySTR = 'The contents of this string have no importance';
sMySTR.search('\Q' + sNeedle + '\E');

Result is "invalid quantifier". This happens because 2 or more special characters are 'touching' each other, because:

var sNeedle = '*Dont touch me*Stars!*Dont touch me*';
var sMySTR = 'The contents of this string have no importance';
sMySTR.search('\Q' + sNeedle + '\E');

Will work OK.

I know I could make a function escapeAllBadChars(sInStr) and just add double slashes before every possible special regex character, but I'm wondering if there is a simpler way to do it?

Answer 1

\Q...\E doesn't work in JavaScript (at least, they don't escape anything...) as you can see:

var s = "*";
print(s.search(/\Q*\E/));
print(s.search(/\*/));

produces:

-1
0

as you can see on Ideone.

The following chars need to be escaped:

• (
• )
• [
• {
• *
• +
• .
• $
• ^
• \
• |
• ?

So, something like this would do:

function quote(regex) {
  return regex.replace(/([()[{*+.$^\\|?])/g, '\\$1');
}

No, ] and } don't need to be escaped: they have no special meaning, only their opening counter parts.

Note that when using a literal regex, /.../, you also need to escape the / char. However, / is not a regex meta character: when using it in a RegExp object, it doesn't need an escape.