Trim white spaces in both Object key and value recursively

问题

How do you trim white spaces in both the keys and values in a JavaScript Object recursively?

I came across one issue in which I was trying to "clean" a user supplied JSON string and send it into my other code for further processing.

Let's say we've got a user supplied JSON string whose property key and value are of type "string". However, what's problematic in this case is that the keys and values are not as clean as desired. Say a { " key_with_leading_n_trailing_spaces ": " my_value_with_leading_spaces" }.

In this case, it can easily cause issue with your brilliantly written JavaScript program trying to make use of such data(or should we call it dirty data?) because when your code is trying to get the value out of this JSON object, not only the key is not matching but also the value can not be matched. I have looked around google and found a few tips but there is not one cure that cure it all.

Given this JSON with lots of white spaces in keys and values.

var badJson = {
  "  some-key   ": "    let it go    ",
  "  mypuppy     ": "    donrio   ",
  "   age  ": "   12.3",
  "  children      ": [
    { 
      "   color": " yellow",
      "name    ": "    alice"
    },    { 
      "   color": " silver        ",
      "name    ": "    bruce"
    },    { 
      "   color": " brown       ",
      "     name    ": "    francis"
    },    { 
      "   color": " red",
      "      name    ": "    york"
    },

  ],
  "     house": [
    {
      "   name": "    mylovelyhouse     ",
      " address      " : { "number" : 2343, "road    "  : "   boardway", "city      " : "   Lexiton   "}
    }
  ]

};

So this is what I came up with ( with help of using lodash.js):

//I made this function to "recursively" hunt down keys that may 
//contain leading and trailing white spaces
function trimKeys(targetObj) {

  _.forEach(targetObj, function(value, key) {

      if(_.isString(key)){
        var newKey = key.trim();
        if (newKey !== key) {
            targetObj[newKey] = value;
            delete targetObj[key];
        }

        if(_.isArray(targetObj[newKey]) || _.isObject(targetObj[newKey])){
            trimKeys(targetObj[newKey]);
        }
      }else{

        if(_.isArray(targetObj[key]) || _.isObject(targetObj[key])){
            trimKeys(targetObj[key]);
        }
      }
   });

}

//I stringify this is just to show it in a bad state
var badJson = JSON.stringify(badJson);

console.log(badJson);

//now it is partially fixed with value of string type trimed
badJson = JSON.parse(badJson,function(key,value){
    if(typeof value === 'string'){
        return value.trim();
    }
    return value;
});

trimKeys(badJson);

console.log(JSON.stringify(badJson));

Note here : I did this in a 1, 2 steps because I could not find a better one shot to deal it all solution. If there is issue in my code or anything better, please do share with us.

Thanks!

回答1:

You can clean up the property names and attributes using Object.keys to get an array of the keys, then Array.prototype.reduce to iterate over the keys and create a new object with trimmed keys and values. The function needs to be recursive so that it also trims nested Objects and Arrays.

Note that it only deals with plain Arrays and Objects, if you want to deal with other types of object, the call to reduce needs to be more sophisticated to determine the type of object (e.g. a suitably clever version of new obj.constructor()).

function trimObj(obj) {
  if (!Array.isArray(obj) && typeof obj != 'object') return obj;
  return Object.keys(obj).reduce(function(acc, key) {
    acc[key.trim()] = typeof obj[key] == 'string'? obj[key].trim() : trimObj(obj[key]);
    return acc;
  }, Array.isArray(obj)? []:{});
}

回答2:

You can just stringify it, string replace, and reparse it

JSON.parse(JSON.stringify(badJson).replace(/"\s+|\s+"/g,'"'))

回答3:

epascarello's answer above plus some unit tests (just for me to be sure):

function trimAllFieldsInObjectAndChildren(o: any) {
  return JSON.parse(JSON.stringify(o).replace(/"\s+|\s+"/g, '"'));
}

import * as _ from 'lodash';
assert.true(_.isEqual(trimAllFieldsInObjectAndChildren(' bob '), 'bob'));
assert.true(_.isEqual(trimAllFieldsInObjectAndChildren('2 '), '2'));
assert.true(_.isEqual(trimAllFieldsInObjectAndChildren(['2 ', ' bob ']), ['2', 'bob']));
assert.true(_.isEqual(trimAllFieldsInObjectAndChildren({'b ': ' bob '}), {'b': 'bob'}));
assert.true(_.isEqual(trimAllFieldsInObjectAndChildren({'b ': ' bob ', 'c': 5, d: true }), {'b': 'bob', 'c': 5, d: true}));
assert.true(_.isEqual(trimAllFieldsInObjectAndChildren({'b ': ' bob ', 'c': {' d': 'alica c c '}}), {'b': 'bob', 'c': {'d': 'alica c c'}}));
assert.true(_.isEqual(trimAllFieldsInObjectAndChildren({'a ': ' bob ', 'b': {'c ': {'d': 'e '}}}), {'a': 'bob', 'b': {'c': {'d': 'e'}}}));
assert.true(_.isEqual(trimAllFieldsInObjectAndChildren({'a ': ' bob ', 'b': [{'c ': {'d': 'e '}}, {' f ': ' g ' }]}), {'a': 'bob', 'b': [{'c': {'d': 'e'}}, {'f': 'g' }]}));

回答4:

Similar to epascarello's answer. This is what I did :

import java.util.regex.Matcher;
import java.util.regex.Pattern;

........

public String trimWhiteSpaceAroundBoundary(String inputJson) {
    String result;
    final String regex = "\"\\s+|\\s+\"";
    final Pattern pattern = Pattern.compile(regex);
    final Matcher matcher = pattern.matcher(inputJson.trim());
    // replacing the pattern twice to cover the edge case of extra white space around ','
    result = pattern.matcher(matcher.replaceAll("\"")).replaceAll("\"");
    return result;
}

Test cases

assertEquals("\"2\"", trimWhiteSpace("\" 2 \""));
assertEquals("2", trimWhiteSpace(" 2 "));
assertEquals("{   }", trimWhiteSpace("   {   }   "));
assertEquals("\"bob\"", trimWhiteSpace("\" bob \""));
assertEquals("[\"2\",\"bob\"]", trimWhiteSpace("[\"  2  \",  \"  bob  \"]"));
assertEquals("{\"b\":\"bob\",\"c c\": 5,\"d\": true }",
              trimWhiteSpace("{\"b \": \" bob \", \"c c\": 5, \"d\": true }"));

来源：https://stackoverflow.com/questions/33510625/trim-white-spaces-in-both-object-key-and-value-recursively