Just for future Googlers (or not necessarily 'Googlers'):
All of solutions mentioned above are wonderful, however, RegExp might be awfully bad thing to use in a situation like that.
So, yes, you might use some of the options proposed or even write something primitive yet useful like:
const strToNum = str => {
//Find 1-3 digits followed by exactly 3 digits & a comma or end of string
let regx = /(\d{1,3})(\d{3}(?:,|$))/;
let currStr;
do {
currStr = (currStr || str.split(`.`)[0])
.replace( regx, `$1,$2`)
} while (currStr.match(regx)) //Stop when there's no match & null's returned
return ( str.split(`.`)[1] ) ?
currStr.concat(`.`, str.split(`.`)[1]) :
currStr;
};
strToNum(`123`) // => 123
strToNum(`123456`) // => 123,456
strToNum(`-1234567.0987`) // => -1,234,567.0987
The regexp that's used here is fairly simple and the loop will go precisely the number of times it takes to get the job done.
And you might optimize it far better, "DRYify" code & so on.
Yet,
(-1234567.0987).toLocaleString();
(in most situations) would be a far better choice.
The point is not in the speed of execution or in cross-browser compatibility.
In situations when you'd like to show the resulting number to user, .toLocaleString() method gives you superpower to speak the same language with the user of your website or app (whatever her/his language is).
This method according to ECMAScript documentation was introduced in 1999, and I believe that the reason for that was the hope that the Internet at some point will connect people all around the world, so, some "internalization" tools were needed.
Today the Internet does connect all of us, so, it is important to remember that the world is a way more complex that we might imagine & that (/almost) all of us are here, in the Internet.
Obviously, considering the diversity of people, it is impossible to guarantee perfect UX for everybody because we speak different languages, value different things, etc. And exactly because of this, it is even more important to try to localize things as much as it's possible.
So, considering that there're some particular standards for representation of date, time, numbers, etc. & that we have a tool to display those things in the format preferred by the final user, isn't that rare and almost irresponsible not to use that tool (especially in situations when we want to display this data to the user)?
For me, using RegExp instead of .toLocaleString() in situation like that sounds a little bit like creating a clock app with JavaScript & hard-coding it in such a way so it'll display Prague time only (which would be quite useless for people who don't live in Prague) even though the default behaviour of
new Date();
is to return the data according to final user's clock.
toLocaleString
appears to work fine on Safari, and I'm seeing consistent results between Chrome, Safari, and Firefox. – Roast