is it possible to use regex in js to replace several different chars?

Question

I need to replace all accented char in a string by it's unaccented version, for sorting. I found how to match the accented ones, but is it possible to use a regex to replace each one? I mean:

var re = /ùÙüÜäàáëèéïìíöòóüùúÄÀÁËÈÉÏÌÍÖÒÓÜÚñÑ/g;
var str = "ùÙüÜäàáëèéïìíöòóüùúÄÀÁËÈÉÏÌÍÖÒÓÜÚñÑ";
var newstr = str.replace(re, 'M');
console.log(newstr);

this prints 'M' but I need :'uUuUaaaeeeiiiooouuuAAAEEEIIIOOOUUnN'

Is this possible? thanks

Answer 1

You need to use character classes.

var re = /[ùÙüÜäàáëèéïìíöòóüùúÄÀÁËÈÉÏÌÍÖÒÓÜÚñÑ]/g;

Then, you can pass a function as a second argument to the replace function. This function shall contain the conversion logic. A simple way would be to use a conversion map.

E.g.

var re = /[ùÙüÜäàáëèéïìíöòóüùúÄÀÁËÈÉÏÌÍÖÒÓÜÚñÑ]/g;

//incomplete but you get the idea
var conversionMap = {
    'ù': 'u',
    'Ù': 'U',
    'ü': 'u',
    'Ü': 'U',
    'ä': 'a'
};

"ùÙüÜäàáëèéïìíöòóüùúÄÀÁËÈÉÏÌÍÖÒÓÜÚñÑ".replace(re, function (c) {
    return conversionMap[c] || c;
}); //uUuUaàáëèéïìíöòóuuúÄÀÁËÈÉÏÌÍÖÒÓUÚñÑ

FIDDLE

Answer 2

http://jsfiddle.net/Victornpb/YPtaN/4

var deaccentuate = (function(){

    var accent = "ùÙüÜäàáëèéïìíöòóüùúÄÀÁËÈÉÏÌÍÖÒÓÜÚñÑ",
        latin  = "uUuUaaaeeeiiiooouuuAAAEEEIIIOOOUUnN".split("");

    var re = new RegExp("["+accent+"]", "g");

    return function(str){
        return str.replace(re, function(c){
            return latin[accent.indexOf(c)]; }
        );
    } 
})();

deaccentuate("Olá, como estás?"); //Ola, como estas?

Benchmark

I realized a benchmark test with a 2KB text and my function was faster than other answers, reaching 59000 Ops/sec

http://jsperf.com/deaccentuate

Answer 3

This is fairly verbose, in order to be readable. (Well, to each their own, anyway.)

var deaccentuate = (function() {
  var conversion =
      { 'a' : /[äàá]/g
      , 'e' : /[ëèé]/g
      , 'i' : /[ïìí]/g
      , 'o' : /[öòó]/g
      , 'u' : /[üùú]/g
      , 'n' : /ñ/g
      , 'A' : /[ÄÀÁ]/g
      , 'E' : /[ËÈÉ]/g
      , 'I' : /[ÏÌÍ]/g
      , 'O' : /[ÖÒÓ]/g
      , 'U' : /[ÜÙÚ]/g
      , 'N' : /Ñ/g
      }

  return function(str) {
    return Object.keys(conversion).reduce(function(str, c) {
      return str.replace(conversion[c], c)
    }, str)
  }
}())

Usage: (http://jsbin.com/UFEbuho/1/)

var input = "ùÙüÜäàáëèéïìíöòóüùúÄÀÁËÈÉÏÌÍÖÒÓÜÚñÑ"

console.log(deaccentuate(input))

The idea is to loop over the keys of the conversion table and replace anything that matches the pattern of that key to the key itself. This is certainly not the most efficient way to do this, but unless the input strings are fairly long it shouldn't matter much.

Answer 4

I can't think about an easier way to efficiently remove all diacritics from a string than using this amazing solution.

See it in action:

var str = "ùÙüÜäàáëèéïìíöòóüùúÄÀÁËÈÉÏÌÍÖÒÓÜÚñÑ";

var str_norm = str.normalize('NFD').replace(/[\u0300-\u036f]/g, '');
console.log(str_norm);