Regex to split string into separate keys

Question

Working on a bit of an interesting project and I'm having a little trouble writing regex for this.

Using regex and Javascript, I want to split my input string into an array of "keys". Strings that have an invalid key should return 0 keys (null, empty array, whatever is fine).

The "spec"

<input string>      => <matches>

# named keys
a                   => ['a']
b                   => ['b']
foo[bar]            => ['foo', 'bar']

# numeric keys are ok too
a[0]                => ['a', '0']
b[1]                => ['b', '1']

# 'push' syntax is also valid
push[]              => ['push', '']
hello[]             => ['hello', '']

# complex!
complex[named][0][] => ['complex', 'named', '0', '']

Invalid examples

# named keys must start with [a-zA-Z]
_invalid            => null
3foo[abc]           => null

# named keys can include a number as long as they start with a letter
foo3valid[this_is_valid_2] => ['foo3valid', 'this_is_valid_2']

What I have so far

var regex = /^(?:\[?([a-zA-Z][a-zA-Z0-9_]*|[a-zA-Z0-9_]*)\]?)+$/g;
var keys = [];

myInput.replace(regex, function(match,capture,pos,source){
    keys.push(capture);
});

console.log(myInput, keys);

How it fails

My regex is matching only the last "key" E.g.,

# should be ['a', 'b', 'c']
a[b][c]   => ['c']

Tinker on jsFiddle

jsfiddle

@MikePartridge, I had a false positive result in a different test. You're right, the regex above only captures the last key. (I updated the question) — maček
– maček, Commented Dec 6, 2011 at 18:39

Rob W · Accepted Answer · 2011-12-06 18:58:23Z

1

I have taken another approach to get the desired results:

Test whether the given phrase is valid.
If valid, use a global match to create an array.

Code: Fiddle: http://jsfiddle.net/MVJZc/2/

///Example
var myInput = "a[b][][c]";

var validate = /^[a-zA-Z][a-zA-Z0-9_]*(?:\[(?:\d*|[a-zA-Z0-9_]+)\])*$/,
    pattern = /[a-zA-Z0-9_]+|(?=\[\])/g,
    //(?=\[\]) is used to create an empty entry for []
    keys = null;
if (validate.test(myInput)) {
    keys = myInput.match(pattern);
}

Note: If you want to extend the pattern, e.g. to include the $ character, modify the first RegEx. The second RegEx (pattern) should at least contain all valid characters.

edited Dec 6, 2011 at 18:58

answered Dec 6, 2011 at 18:52

Rob W

350k87 gold badges811 silver badges683 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

maček Over a year ago

This is getting close, but some inputs aren't passing. See results here: jsfiddle.net/MVJZc/3

Rob W Over a year ago

@macek Use my fiddle. My very first answer (which you've used for your fiddle) was missing the + sign at the validate pattern. jsfiddle.net/MVJZc/2

Rob W Over a year ago

By the way, use /^[a-zA-Z][a-zA-Z0-9_]*(?:\[(?:\d*|[a-zA-Z][a-zA-Z0-9_]+)\])*$/ if a [...] may only start with a digit if it's only consisting of digits, or /^[a-zA-Z][a-zA-Z0-9_]*(?:\[[a-zA-Z0-9_]+\])*$/ if b[3ab] is also valid.

Joseph Marikle · Accepted Answer · 2011-12-06 18:38:20Z

1

would something like this work?

var input = "a[b][c]";
input = input.replace(/[\[\]]$/,"").replace(/[\]\[]+/g,"|");
var args = input.split("|");

And here's a demo: http://jsfiddle.net/vZBtj/

answered Dec 6, 2011 at 18:38

Joseph Marikle

78.8k18 gold badges114 silver badges130 bronze badges

2 Comments

maček Over a year ago

This does not validate each key. It's only validating the first one. E.g., foo[_invalid] and foo[bar][ fails2] should not return keys.

Joseph Marikle Over a year ago

@macek I see... I would go back and fix it but it seems you found an answer already. :P

Upgradingdave · Accepted Answer · 2011-12-06 19:23:47Z

Here's another possibility, definitely not as succinct as Rob W's, but get's the job done ;-)

function parseThing(aStr){
    var res = [];
    success = false;
    if(aStr.indexOf('[') > 0){
        var idx = aStr.indexOf('[');
        var first = aStr.substring(0, idx);
        if(first.match(/\w/)){
            success = true;
            res.push(first);

            //now parse stuff inside brackets
            var rest = aStr.substring(idx, aStr.length-1);
            var arr = rest.split("]");
            for(i in arr){
                res.push(arr[i].replace("[", ""));
            }
        }
    } else if(aStr.match(/\w/)){
        res.push(aStr);
    }
    return res;
}

console.log(parseThing("a[a][b]"));
console.log(parseThing("a"));
console.log(parseThing("b"));
console.log(parseThing("foo[bar] "));
console.log(parseThing("a[0]"));
console.log(parseThing("b[1]"));
console.log(parseThing("complex[named][0][]"));

Collectives™ on Stack Overflow

Regex to split string into separate keys

The "spec"

Invalid examples

What I have so far

How it fails

Tinker on jsFiddle

3 Answers 3

3 Comments

2 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

The "spec"

Invalid examples

What I have so far

How it fails

Tinker on jsFiddle

3 Answers 3

3 Comments

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related