What is wrong with this regex for comma-separated string?

Question

I'm trying to convert a comma or semicolon separated string into an array, but handling before conversion several omission from users (double commas, space after comma or semicolon). In the example:

myStr = "item1, item2,,, this is item3; item4 , "

I want to get:

myArray = ["item1", "item2", "this is item3", "item4"]

My actual regex is:

myArray = myStr.split(/[(,\s);,]+/);

But, although I'm adding parenthesis to the combination comma-space (,\s) the regex are catching the spaces inside the third item. Any advice what could be wrong with regex?

Possible duplicate of [Convert comma separated string to array](http://stackoverflow.com/questions/2858121/convert-comma-separated-string-to-array) — , Jun 01 '16 at 01:25
Not a duplicate; this question also has empty item elimination and isn't just commas. — Jacob, Jun 01 '16 at 01:26
@Mike, is not the same. I want to trim spaces, not using them as separators. — Ruben, Jun 01 '16 at 01:59

score 2 · Accepted Answer · edited Jun 01 '16 at 02:29

2

Try this (edited):

var myStr = "item1, item2,,, this is item3; item4 , ";
var myArray = myStr.split(/\s*[,;]+\s*/).filter(Boolean);
alert(myArray);

edited Jun 01 '16 at 02:29

Ruben

816
1
8
21

answered Jun 01 '16 at 01:29

Jaumzera

2,305
1
30
44

Doesn't handle the trimming as mentioned in the question. – Jacob Jun 01 '16 at 01:41
Well, I fixed the code the way you ask. I think it would attend you. Take a look, please. – Jaumzera Jun 01 '16 at 02:00
@Jaumzera, its almost Ok, just the split let an empty item ("") at the end of array. – Ruben Jun 01 '16 at 02:02
Well, I quit... I can't figure out a way to do that in a elegant and readable way. – Jaumzera Jun 01 '16 at 02:12
Thanx @Ruben... it fits perfectly. – Jaumzera Jun 01 '16 at 02:31
Thanks to @Jacob contribution, the `filter(Boolean)` avoid the empty item left at the end of array =) – Ruben Jun 01 '16 at 02:37

Jacob · Answer 2 · 2016-06-01T01:23:35.000

1

[] only contains single characters or character classes and only matches one character at a time. Therefore, [(,\s);,] really means any of these characters are matched: (),; or whitespace.

There are easier approaches than regex. Try this:

myArray = myStr
  .split(/[;,]/)                 
  .map(function (str) { 
    return str.replace(/(^\s+)|(\s+$)/g, ''); // Trim whitespace
  }) 
  .filter(Boolean);              // Filter away the empty strings

edited Jun 01 '16 at 01:23

answered Jun 01 '16 at 01:17

Jacob

77,566
24
149
228

Wouldn't this return `["item1", "item2", "this", "is", "item3", "item4", ""]`? – choz Jun 01 '16 at 01:21
@choz, The `.filter(Boolean)` removes any empty entries. – Jacob Jun 01 '16 at 01:25
Thanks @Jacob. Your code works great. Anyway I give the Ok to the answer using regex. I used your `filter(Boolean)` trick to delete the empty item in array which regex can't delete. – Ruben Jun 01 '16 at 02:31

score 1 · Answer 3 · answered Jun 01 '16 at 01:22

1

I think you need to ensure you always match at least one non-white-space character. Try this:

myArray = myStr.split(/[\s]*[,;][,;\s]*/);

answered Jun 01 '16 at 01:22

Alastair Brown

1,598
8
12

Almost done, just this regex let an empty item ("") at the end of the array. – Ruben Jun 01 '16 at 01:55

mike510a · Answer 4 · 2016-06-01T02:26:53.280

1

var myArray = myStr.split(/(?: , | ; | ,| ;|, |; |;|,)+/);  

 Then do a quick trim() and verify if the item has content other than whitespace

Would effectively split a string using , or ; as the separator

var myStr = "item1, item2,,, this is item3; item4 , ";
var myArray = myStr.split(/(?: , | ; | ,| ;|, |; |;|,)+/); 
var newArray = [];
for (var i=0; i<myArray.length; i++) {
 var checkArray = myArray[i].trim();
 if (checkArray.length > 0) { newArray.push(myArray[i]); }
}
myArray = newArray;

console.log(myArray);

ANSWER 2

The best answer I could come up with was one I made by combining @Jacob's answer and my answer above.

This will effectively Accomplish the job in the fastest way possible

var myStr = "item1, item2,,, this is item3; item4 , ";
var myArray = myStr.split(/(?:,|;)+/).map(function (str) {
  return str.replace(/(^\s+)|(\s+$)/g, '');
 }).filter(Boolean);
console.log(myArray);

edited Jun 01 '16 at 02:26

answered Jun 01 '16 at 01:27

mike510a

2,102
1
11
27

Doesn't trim or eliminate empty entries as in the question. – Jacob Jun 01 '16 at 01:41
Also, why replace a perfectly good character class with a less-efficient, less-readable alternation? – Alan Moore Jun 01 '16 at 01:50
@Alan Well, mainly because of the requirements outlined in the OP where he wanted to match against more than one character sometimes – mike510a Jun 01 '16 at 02:17
@Jacob sure it does. – mike510a Jun 01 '16 at 02:18

What is wrong with this regex for comma-separated string?

4 Answers4