Javascript：用逗号分隔字符串，括号内除外

Question

Given string in the form:

'"abc",ab(),c(d(),e()),f(g(),zyx),h(123)'

How can I split it to get the below array format:

abc
ab()
c(d(),e())
f(g(),zyx)
h(123)

I have tried normal javascript split, however it doesn't work as desired. Trying Regular Expression but not yet successful.

Answer 1

You can keep track of the parentheses, and add those expressions when the left and right parens equalize.

For example-

function splitNoParen(s){
    var left= 0, right= 0, A= [], 
    M= s.match(/([^()]+)|([()])/g), L= M.length, next, str= '';
    for(var i= 0; i<L; i++){
        next= M[i];
        if(next=== '(')++left;
        else if(next=== ')')++right;
        if(left!== 0){
            str+= next;
            if(left=== right){
                A[A.length-1]+=str;
                left= right= 0;
                str= '';
            }
        }
        else A=A.concat(next.match(/([^,]+)/g));
    }
    return A;
}

var s1= '"abc",ab(),c(d(),e()),f(g(),zyx),h(123)';
splitNoParen(s1).join('\n');

/*  returned value: (String)
"abc"
ab()
c(d(),e())
f(g(),zyx)
h(123)
*/

Answer 2

This might be not the best or more refined solution, and also maybe won't fit every single possibility, but based on your example it works:

var data = '"abc",ab(),c(d(),e()),f(g(),zyx),h(123)';
// Create a preResult splitting the commas.
var preResult = data.replace(/"/g, '').split(',');
// Create an empty result.
var result = [];

for (var i = 0; i < preResult.length; i++) {
    // Check on every preResult if the number of parentheses match.
    // Opening ones...
    var opening = preResult[i].match(/\(/g) || 0;
    // Closing ones...
    var closing = preResult[i].match(/\)/g) || 0;

    if (opening != 0 &&
        closing != 0 &&
        opening.length != closing.length) {
        // If the current item contains a different number of opening
        // and closing parentheses, merge it with the next adding a 
        // comma in between.
        result.push(preResult[i] + ',' + preResult[i + 1]);
        i++;
    } else {
        // Leave it as it is.
        result.push(preResult[i]);
    }
}

Demo

Answer 3

For future reference, here's another approach to top-level splitting, using string.replace as a control flow operator:

function psplit(s) {
  var depth = 0, seg = 0, rv = [];
  s.replace(/[^(),]*([)]*)([(]*)(,)?/g,
            function (m, cls, opn, com, off, s) {
    depth += opn.length - cls.length;
    var newseg = off + m.length;
    if (!depth && com) {
      rv.push(s.substring(seg, newseg - 1));
      seg = newseg;
    }
    return m;
  });
  rv.push(s.substring(seg));
  return rv;
}

console.log(psplit('abc,ab(),c(d(),e()),f(g(),zyx),h(123)'))

["abc", "ab()", "c(d(),e())", "f(g(),zyx)", "h(123)"]

Getting it to handle quotes as well would not be too complicated, but at some point you need to decide to use a real parser such as jison , and I suspect that would be the point. In any event, there's not enough detail in the question to know what the desired handling of double quotes is.

Answer 4

You can't use .split for this, but instead you'll have to write a small parser like this:

 function splitNoParen(s){ let results = []; let next; let str = ''; let left = 0, right = 0; function keepResult() { results.push(str); str = ''; } for(var i = 0; i<s.length; i++) { switch(s[i]) { case ',': if((left === right)) { keepResult(); left = right = 0; } else { str += s[i]; } break; case '(': left++; str += s[i]; break; case ')': right++; str += s[i]; break; default: str += s[i]; } } keepResult(); return results; } var s1= '"abc",ab(),c(d(),e()),f(g(),zyx),h(123)'; console.log(splitNoParen(s1).join('\n')); var s2='cats,(my-foo)-bar,baz'; console.log(splitNoParen(s2).join('\n'));

Answer 5

Had a similar issue and existing solutions were hard to generalize. So here's another parser that's a bit more readable and easier to extend to your personal needs. It'll also work with curly braces, brackets, normal braces, and strings of any type. License is MIT.

export function parse_arguments(input:string, split_by = ",")
{
    // Some symbols can be nested, like braces, and must be counted
    const state = {"{":0,"[":0,"(":0};
    // Some cannot be nested, and just flip a flag
    const state_singlo = {"\"":false,"'":false,"\`":false}

    // Nestable symbols come in sets, usually in pairs.
    // These sets increase or decrease the state, depending on the symbol.
    const pairs : Record<string,[keyof typeof state,number]> = {
        "{":["{",1],
        "}":["{",-1],
        "[":["[",1],
        "]":["[",-1],
        "(":["(",1],
        ")":["(",-1]
    }

    let start = 0;
    let results = [];
    let length = input.length;
    for(let i = 0; i < length; ++i)
    {
        let char = input[i];

        // Backslash escapes the next character. We directly skip 2 characters by incrementing i one extra time.
        if(char === "\\")
        {
            i++;
            continue;
        }
        // If it's a paired symbol, increase or decrease the state based on our "pairs" constant.
        if(char in pairs)
        {
            let [key,value] = pairs[char];
            state[key] += value;
        }
        // If the symbol exists in the single/not nested state object, flip the corresponding state flag.
        else if(char in state_singlo)
        {
            state_singlo[char as keyof typeof state_singlo] = !state_singlo[char as keyof typeof state_singlo];
        }
        // If it's our split symbol, push the 
        else if(char === split_by)
        {
            if(Object.entries(state).every(([k,v])=>v == 0) && Object.entries(state_singlo).every(([k,v])=>!v))
            {
                results.push(input.substring(start,i))
                start = i+1;
            }
        }
    }

    // Add the last segment if the string didn't end in the split_by symbol, otherwise add an empty string
    if(start < input.length)
        results.push(input.substring(start,input.length))
    else
        results.push("");

    return results;
}

Answer 6

With this regex, it makes the job:

 const regex = /,(?;[^(]*\))/g, const str = '"abc",ab(),c(d(),e()),f(g(),zyx);h(123)'. const result = str;split(regex). console;log(result);

Answer 7

Javascript

var str='"abc",ab(),c(d(),e()),f(g(),zyx),h(123)'
str.split('"').toString().split(',').filter(Boolean);

this should work