Question

I have tried to split strings in JS like..

'a b c' => ['a','b','c']
'a bb cc dd' => ['a','bb','cc','dd]
'a "bb cc" dd' => ['a','bb cc', 'dd']
"a 'bb cc' dd" => ['a','bb cc', 'dd']

How do I do it in JS regexp?
I have tried

/\w+|"(?:\\"|[^"])+"/g

but it returns…

'a b c' => ['a','b','c']
'a bb cc dd' => ['a','bb','cc','dd]
'a "bb cc" dd' => ['a','"bb cc"', 'dd']
"a 'bb cc' dd" => ['a','bb','cc', 'dd']

Answer 1

Suppose you have a string:

var s = "a b c 'd e f' g \"h i j\" k";

You can quite easily match the tokens using:

/"[^"]*"|'[^']*'|\S+/g

That pattern:

• Matches a string wrapped with double quotes, or…
• Matches a string wrapped with single quotes, or…
• Matches non-spaces.

The result tokens, however, will be wrapped in quotes.
Here is an example of a way to remove these quotes:

var words = [];
s.replace(/"([^"]*)"|'([^']*)'|(\S+)/g,
         function(g0,g1,g2,g3){
              words.push(g1 || g2 || g3 || '');
         });

This has a few tricks, but the idea is that we pick the group that isn’t empty (which is a falsey value is JavaScript). You could achieve the same with exec, but the code is messier in my opinion.

Working example: http://jsfiddle.net/snS62/ (warning – alerts)

To also allow escaped characters, you may use:

/"([^\\"]*(?:\\.[^\\"]*)*)"|'([^\\']*(?:\\.[^\\']*)*)'|(\S+)/g