How to split a long regular expression into multip

Question

I have a very long regular expression, which I wish to split into multiple lines in my JavaScript code to keep each line length 80 characters according to JSLint rules. It's just better for reading, I think. Here's pattern sample:

var pattern = /^(([^<>()[\]\\.,;:\s@\"]+(\.[^<>()[\]\\.,;:\s@\"]+)*)|(\".+\"))@((\[[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\])|(([a-zA-Z\-0-9]+\.)+[a-zA-Z]{2,}))$/;

Answer 1

You can simply use string operation.

var pattenString = "^(([^<>()[\]\\.,;:\s@\"]+(\.[^<>()[\]\\.,;:\s@\"]+)*)|"+
"(\".+\"))@((\[[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\])|"+
"(([a-zA-Z\-0-9]+\.)+[a-zA-Z]{2,}))$";
var patten = new RegExp(pattenString);

Answer 2

The regex above is missing some black slashes which isn't working properly. So, I edited the regex. Please consider this regex which works 99.99% for email validation.

let EMAIL_REGEXP = 
new RegExp (['^(([^<>()[\\]\\\.,;:\\s@\"]+(\\.[^<>()\\[\\]\\\.,;:\\s@\"]+)*)',
                    '|(".+"))@((\\[[0-9]{1,3}\\.[0-9]{1,3}\\.[0-9]{1,3}\\.',
                    '[0-9]{1,3}\])|(([a-zA-Z\\-0-9]+\\.)+',
                    '[a-zA-Z]{2,}))$'].join(''));

Answer 3

You could convert it to a string and create the expression by calling new RegExp():

var myRE = new RegExp (['^(([^<>()[\]\\.,;:\\s@\"]+(\\.[^<>(),[\]\\.,;:\\s@\"]+)*)',
                        '|(\\".+\\"))@((\\[[0-9]{1,3}\\.[0-9]{1,3}\\.[0-9]{1,3}\\.',
                        '[0-9]{1,3}\])|(([a-zA-Z\-0-9]+\\.)+',
                        '[a-zA-Z]{2,}))$'].join(''));

Notes:

1. when converting the expression literal to a string you need to escape all backslashes as backslashes are consumed when evaluating a string literal. (See Kayo's comment for more detail.)

2. RegExp accepts modifiers as a second parameter

   /regex/g => new RegExp('regex', 'g')

[Addition ES20xx (tagged template)]

In ES20xx you can use tagged templates. See the snippet.

Note:

• Disadvantage here is that you can't use plain whitespace in the regular expression string (always use \s, \s+, \s{1,x}, \t, \n etc).

(() => {
  const createRegExp = (str, opts) => 
    new RegExp(str.raw[0].replace(/\s/gm, ""), opts || "");
  const yourRE = createRegExp`
    ^(([^<>()[\]\\.,;:\s@\"]+(\.[^<>()[\]\\.,;:\s@\"]+)*)|
    (\".+\"))@((\[[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\.[0-9]{1,3}\])|
    (([a-zA-Z\-0-9]+\.)+[a-zA-Z]{2,}))$`;
  console.log(yourRE);
  const anotherLongRE = createRegExp`
    (\byyyy\b)|(\bm\b)|(\bd\b)|(\bh\b)|(\bmi\b)|(\bs\b)|(\bms\b)|
    (\bwd\b)|(\bmm\b)|(\bdd\b)|(\bhh\b)|(\bMI\b)|(\bS\b)|(\bMS\b)|
    (\bM\b)|(\bMM\b)|(\bdow\b)|(\bDOW\b)
    ${"gi"}`;
  console.log(anotherLongRE);
})();

Answer 4

Using strings in new RegExp is awkward because you must escape all the backslashes. You may write smaller regexes and concatenate them.

Let's split this regex

/^foo(.*)\bar$/

We will use a function to make things more beautiful later

function multilineRegExp(regs, options) {
    return new RegExp(regs.map(
        function(reg){ return reg.source; }
    ).join(''), options);
}

And now let's rock

var r = multilineRegExp([
     /^foo/,  // we can add comments too
     /(.*)/,
     /\bar$/
]);

Since it has a cost, try to build the real regex just once and then use that.

Answer 5

Personally, I'd go for a less complicated regex:

/\S+@\S+\.\S+/

Sure, it is less accurate than your current pattern, but what are you trying to accomplish? Are you trying to catch accidental errors your users might enter, or are you worried that your users might try to enter invalid addresses? If it's the first, I'd go for an easier pattern. If it's the latter, some verification by responding to an e-mail sent to that address might be a better option.

However, if you want to use your current pattern, it would be (IMO) easier to read (and maintain!) by building it from smaller sub-patterns, like this:

var box1 = "([^<>()[\]\\\\.,;:\s@\"]+(\\.[^<>()[\\]\\\\.,;:\s@\"]+)*)";
var box2 = "(\".+\")";

var host1 = "(\\[[0-9]{1,3}\\.[0-9]{1,3}\\.[0-9]{1,3}\\.[0-9]{1,3}\\])";
var host2 = "(([a-zA-Z\-0-9]+\\.)+[a-zA-Z]{2,})";

var regex = new RegExp("^(" + box1 + "|" + box2 + ")@(" + host1 + "|" + host2 + ")$");

Answer 6

Extending @KooiInc answer, you can avoid manually escaping every special character by using the source property of the RegExp object.

Example:

var urlRegex= new RegExp(''
  + /(?:(?:(https?|ftp):)?\/\/)/.source     // protocol
  + /(?:([^:\n\r]+):([^@\n\r]+)@)?/.source  // user:pass
  + /(?:(?:www\.)?([^\/\n\r]+))/.source     // domain
  + /(\/[^?\n\r]+)?/.source                 // request
  + /(\?[^#\n\r]*)?/.source                 // query
  + /(#?[^\n\r]*)?/.source                  // anchor
);

or if you want to avoid repeating the .source property you can do it using the Array.map() function:

var urlRegex= new RegExp([
  /(?:(?:(https?|ftp):)?\/\/)/      // protocol
  ,/(?:([^:\n\r]+):([^@\n\r]+)@)?/  // user:pass
  ,/(?:(?:www\.)?([^\/\n\r]+))/     // domain
  ,/(\/[^?\n\r]+)?/                 // request
  ,/(\?[^#\n\r]*)?/                 // query
  ,/(#?[^\n\r]*)?/                  // anchor
].map(function(r) {return r.source}).join(''));

In ES6 the map function can be reduced to: .map(r => r.source)