number splitter

README.md

Splits numbers with commas to make them more readable. Thanks to @maettig for being awesome

annotated.js

function(n){
return                                //return
(""+n)                                //the number converted to a sting
.replace(/\B(?=(...)+$)/g,",")        //with a comma inserted every 3 characters
}

index.js

//simple version:
function(n){return(""+n).replace(/\B(?=(...)+$)/g,",")}
//Does not fuck decimals up version:
function f(a,b,c){return b||(c?a+',':a+'').replace(/(\.\d*)|\d(?=(\d{3})+\b)/g,f)}

package.json

{
  "name": "numSplit",

  "description": "Splits numbers with commas to make them more readable.",

  "keywords": [
    "number",
    "comma",
    "math",
    "format"
  ]
}

test.html

<script>
numSplit = function(n){return(""+n).replace(/\B(?=(...)+$)/g,",")}
    
document.write(numSplit(3255325235523632))​​​​​
</script>

You can floor and string-convert a number easier than with the split stuff; use y=''+~~n, the replace will work with "$1,$2", too, it's not necessary to combine those strings.

70 69 bytes. I love a good regular expression. I tried @atk's suggestion but it wont work for numbers that exceed the 32 bit integer limit.

function(a){return(''+a).split('.')[0].replace(/\B(?=(...)+$)/g,',')}

@maettig
O_O I forgot that "g" exists... what a fail!
Meh, who cares about flooring!?:

function(n){return(""+n).replace(/.(?=(...)+$)/g,"$&,")}

(56 bytes)

Now that's a nice one :)

55 bytes. But this works only if you know your input is an integer.

function(a){return(''+a).replace(/\B(?=(...)+$)/g,',')}

I would love to spend the remaining bytes to make this work with every input including strings like "1234 people are spending 1234567.89 USD each 1.2345 days". The following version (58 bytes) almost works except for numbers with fractions. It outputs "1.2,345". I'm not sure how to fix this.

function(a){return(''+a).replace(/\B(?=(\d{3})+\b)/g,',')}

It's hard to explain but it works. Edit: Not enough test cases. "12345" becomes "1234,5" because rule number 3 matches first.

1. Find a dot followed by at least four digits (\.\d{3}\B). Insert a comma after the third digit. Note: \B is a "non-boundary" where a word-character (\w) and another word-character met.
2. Or find a digit that is followed by groups of 3 digits (\d(\d{3})+\b).
3. Continue from either 1. or 2. and insert a comma every three digits. Because of the leading \B this only happens if a digit was matched before.

function(a){return(''+a).replace(/(\.|\B)\d{3}\B|\d(?=(\d{3})+\b)/g,'$&,')}

@maettig
Thats pretty cool :D
But imo it's kinda pointless.
Feel free to fork tho.

Making numSplit(1234.5678) work is pointless?

@maettig
Well, it doesn't really make it work, it adds commas to decimal places... I don't think thats how english works. Then again, I might be wrong about that.

try numSplit(125252.2421415) in Firefox => "1252,52.242,141,5"

Correct would be 125,252.2421415

You are right. I missed that in my tests. What about this?

function(a){return(''+a).replace(/\.\d*|\B(?=(\d{3})+\b)/g,function(a){return a||','})}

Merging the two functions is possible but makes the code very ugly. Edit: OK, here it is. This is smaller and more reliable because it does not insert a comma in "U100".

function f(a,b,c){return b?b:c?a+',':(''+a).replace(/(\.\d*)|\d(?=(\d{3})+\b)/g,f)}

@maettig
Sweet!
1 byte smaller:

function f(a,b,c){return b||(c?a+',':a+'').replace(/(\.\d*)|\d(?=(\d{3})+\b)/g,f)}

Yep, that's cool. This will start recursive calls of replace but it will stop immediately because a contains single digits only. When you do a benchmark you will see it's slower but this can be ignored because of the fast regular expression engines nowadays.

I like it!

What about saving 2 bytes with split?

function(n){return''+(''+n).split(/\B(?=(?:...)+$)/)}

123456789..toLocaleString() seems doing the trick.