[javascript] How to parse JSON using Node.js?

How should I parse JSON using Node.js? Is there some module which will validate and parse JSON securely?

It's simple, you can convert JSON to string using JSON.stringify(json_obj), and convert string to JSON using JSON.parse("your json string").

My solution:

var fs = require('fs');
var file = __dirname + '/config.json';

fs.readFile(file, 'utf8', function (err, data) {
    if (err) {
        console.log('Error: ' + err);
        return;
    }

    data = JSON.parse(data);

    console.dir(data);
});

As mentioned in the above answers, We can use JSON.parse() to parse the strings to JSON But before parsing, be sure to parse the correct data or else it might bring your whole application down

it is safe to use it like this

let parsedObj = {}
try {
    parsedObj = JSON.parse(data);
} catch(e) {
    console.log("Cannot parse because data is not is proper json format")
}

NodeJs is a JavaScript based server, so you can do the way you do that in pure JavaScript...

Imagine you have this Json in NodeJs...

var details = '{ "name": "Alireza Dezfoolian", "netWorth": "$0" }';
var obj = JSON.parse(details);

And you can do above to get a parsed version of your json...

Always be sure to use JSON.parse in try catch block as node always throw an Unexpected Error if you have some corrupted data in your json so use this code instead of simple JSON.Parse

try{
     JSON.parse(data)
}
catch(e){
   throw new Error("data is corrupted")
  }

use the JSON object:

JSON.parse(str);

as other answers here have mentioned, you probably want to either require a local json file that you know is safe and present, like a configuration file:

var objectFromRequire = require('path/to/my/config.json'); 

or to use the global JSON object to parse a string value into an object:

var stringContainingJson = '\"json that is obtained from somewhere\"';
var objectFromParse = JSON.parse(stringContainingJson);

note that when you require a file the content of that file is evaluated, which introduces a security risk in case it's not a json file but a js file.

here, i've published a demo where you can see both methods and play with them online (the parsing example is in app.js file - then click on the run button and see the result in the terminal): http://staging1.codefresh.io/labs/api/env/json-parse-example

you can modify the code and see the impact...

you can require .json files.

var parsedJSON = require('./file-name');

For example if you have a config.json file in the same directory as your source code file you would use:

var config = require('./config.json');

or (file extension can be omitted):

var config = require('./config');

note that require is synchronous and only reads the file once, following calls return the result from cache

Also note You should only use this for local files under your absolute control, as it potentially executes any code within the file.

Include the node-fs library.

var fs = require("fs");
var file = JSON.parse(fs.readFileSync("./PATH/data.json", "utf8"));

For more info on 'fs' library , refer the documentation at http://nodejs.org/api/fs.html

Everybody here has told about JSON.parse, so I thought of saying something else. There is a great module Connect with many middleware to make development of apps easier and better. One of the middleware is bodyParser. It parses JSON, html-forms and etc. There is also a specific middleware for JSON parsing only noop.

Take a look at the links above, it might be really helpful to you.

var fs = require('fs');

fs.readFile('ashish.json',{encoding:'utf8'},function(data,err) {

   if(err) 
      throw err;

   else {

   console.log(data.toString());

 }
})

JSON.parse will not ensure safety of json string you are parsing. You should look at a library like json-safe-parse or a similar library.

From json-safe-parse npm page:

JSON.parse is great, but it has one serious flaw in the context of JavaScript: it allows you to override inherited properties. This can become an issue if you are parsing JSON from an untrusted source (eg: a user), and calling functions on it you would expect to exist.

Just to make this as complicated as possible, and bring in as many packages as possible...

const fs = require('fs');
const bluebird = require('bluebird');
const _ = require('lodash');
const readTextFile = _.partial(bluebird.promisify(fs.readFile), _, {encoding:'utf8',flag:'r'});
const readJsonFile = filename => readTextFile(filename).then(JSON.parse);

This lets you do:

var dataPromise = readJsonFile("foo.json");
dataPromise.then(console.log);

Or if you're using async/await:

let data = await readJsonFile("foo.json");

The advantage over just using readFileSync is that your Node server can process other requests while the file is being read off disk.

No further modules need to be required.
Just use
var parsedObj = JSON.parse(yourObj);
I don think there is any security issues regarding this

Just want to complete the answer (as I struggled with it for a while), want to show how to access the json information, this example shows accessing Json Array:

var request = require('request');_x000D_
request('https://server/run?oper=get_groups_joined_by_user_id&user_id=5111298845048832', function (error, response, body) {_x000D_
  if (!error && response.statusCode == 200) {_x000D_
    var jsonArr = JSON.parse(body);_x000D_
    console.log(jsonArr);_x000D_
    console.log("group id:" + jsonArr[0].id);_x000D_
  }_x000D_
})

If you need to parse JSON with Node.js in a secure way (aka: the user can input data, or a public API) I would suggest using secure-json-parse.

The usage is like the default JSON.parse but it will protect your code from:

• prototype poisoning
• and constructor abuse:

const badJson = '{ "a": 5, "b": 6, "__proto__": { "x": 7 }, "constructor": {"prototype": {"bar": "baz"} } }'

const infected = JSON.parse(badJson)
console.log(infected.x) // print undefined

const x = Object.assign({}, infected)
console.log(x.x) // print 7

const sjson = require('secure-json-parse')
console.log(sjson.parse(badJson)) // it will throw by default, you can ignore malicious data also

JSON.parse("your string");

That's all.

I use fs-extra. I like it a lot because -although it supports callbacks- it also supports Promises. So it just enables me to write my code in a much more readable way:

const fs = require('fs-extra');
fs.readJson("path/to/foo.json").then(obj => {
    //Do dome stuff with obj
})
.catch(err => {
    console.error(err);
});

It also has many useful methods which do not come along with the standard fs module and, on top of that, it also bridges the methods from the native fs module and promisifies them.

NOTE: You can still use the native Node.js methods. They are promisified and copied over to fs-extra. See notes on fs.read() & fs.write()

So it's basically all advantages. I hope others find this useful.

If the JSON source file is pretty big, may want to consider the asynchronous route via native async / await approach with Node.js 8.0 as follows

const fs = require('fs')

const fsReadFile = (fileName) => {
    fileName = `${__dirname}/${fileName}`
    return new Promise((resolve, reject) => {
        fs.readFile(fileName, 'utf8', (error, data) => {
            if (!error && data) {
                resolve(data)
            } else {
                reject(error);
            }
        });
    })
}

async function parseJSON(fileName) {
    try {
        return JSON.parse(await fsReadFile(fileName));
    } catch (err) {
        return { Error: `Something has gone wrong: ${err}` };
    }
}

parseJSON('veryBigFile.json')
    .then(res => console.log(res))
    .catch(err => console.log(err))

Use this to be on the safe side

var data = JSON.parse(Buffer.concat(arr).toString());

Leverage Lodash's attempt function to return an error object, which you can handle with the isError function.

// Returns an error object on failure
function parseJSON(jsonString) {
   return _.attempt(JSON.parse.bind(null, jsonString));
}


// Example Usage
var goodJson = '{"id":123}';
var badJson = '{id:123}';
var goodResult = parseJSON(goodJson);
var badResult = parseJSON(badJson);

if (_.isError(goodResult)) {
   console.log('goodResult: handle error');
} else {
   console.log('goodResult: continue processing');
}
// > goodResult: continue processing

if (_.isError(badResult)) {
   console.log('badResult: handle error');
} else {
   console.log('badResult: continue processing');
}
// > badResult: handle error

Using JSON for your configuration with Node.js? Read this and get your configuration skills over 9000...

Note: People claiming that data = require('./data.json'); is a security risk and downvoting people's answers with zealous zeal: You're exactly and completely wrong. Try placing non-JSON in that file... Node will give you an error, exactly like it would if you did the same thing with the much slower and harder to code manual file read and then subsequent JSON.parse(). Please stop spreading misinformation; you're hurting the world, not helping. Node was designed to allow this; it is not a security risk!

Proper applications come in 3+ layers of configuration:

1. Server/Container config
2. Application config
3. (optional) Tenant/Community/Organization config
4. User config

Most developers treat their server and app config as if it can change. It can't. You can layer changes from higher layers on top of each other, but you're modifying base requirements. Some things need to exist! Make your config act like it's immutable, because some of it basically is, just like your source code.

Failing to see that lots of your stuff isn't going to change after startup leads to anti-patterns like littering your config loading with try/catch blocks, and pretending you can continue without your properly setup application. You can't. If you can, that belongs in the community/user config layer, not the server/app config layer. You're just doing it wrong. The optional stuff should be layered on top when the application finishes it's bootstrap.

Stop banging your head against the wall: Your config should be ultra simple.

Take a look at how easy it is to setup something as complex as a protocol-agnostic and datasource-agnostic service framework using a simple json config file and simple app.js file...

container-config.js...

{
    "service": {
        "type"  : "http",
        "name"  : "login",
        "port"  : 8085
    },
    "data": {
        "type"  : "mysql",
        "host"  : "localhost",
        "user"  : "notRoot",
        "pass"  : "oober1337",
        "name"  : "connect"
    }
}

index.js... (the engine that powers everything)

var config      = require('./container-config.json');       // Get our service configuration.
var data        = require(config.data.type);            // Load our data source plugin ('npm install mysql' for mysql).
var service     = require(config.service.type);         // Load our service plugin ('http' is built-in to node).
var processor   = require('./app.js');                  // Load our processor (the code you write).

var connection  = data.createConnection({ host: config.data.host, user: config.data.user, password: config.data.pass, database: config.data.name });
var server      = service.createServer(processor);
connection.connect();
server.listen(config.service.port, function() { console.log("%s service listening on port %s", config.service.type, config.service.port); });

app.js... (the code that powers your protocol-agnostic and data-source agnostic service)

module.exports = function(request, response){
    response.end('Responding to: ' + request.url);
}

Using this pattern, you can now load community and user config stuff on top of your booted app, dev ops is ready to shove your work into a container and scale it. You're read for multitenant. Userland is isolated. You can now separate the concerns of which service protocol you're using, which database type you're using, and just focus on writing good code.

Because you're using layers, you can rely on a single source of truth for everything, at any time (the layered config object), and avoid error checks at every step, worrying about "oh crap, how am I going to make this work without proper config?!?".

Since you don't know that your string is actually valid, I would put it first into a try catch. Also since try catch blocks are not optimized by node, i would put the entire thing into another function:

function tryParseJson(str) {
    try {
        return JSON.parse(str);
    } catch (ex) {
        return null;
    }
}

OR in "async style"

function tryParseJson(str, callback) {
    process.nextTick(function () {
      try {
          callback(null, JSON.parse(str));
      } catch (ex) {
          callback(ex)
      }
    })
}

You can use JSON.parse().

You should be able to use the JSON object on any ECMAScript 5 compatible JavaScript implementation. And V8, upon which Node.js is built is one of them.

Note: If you're using a JSON file to store sensitive information (e.g. passwords), that's the wrong way to do it. See how Heroku does it: https://devcenter.heroku.com/articles/config-vars#setting-up-config-vars-for-a-deployed-application. Find out how your platform does it, and use process.env to retrieve the config vars from within the code.

Parsing a string containing JSON data

var str = '{ "name": "John Doe", "age": 42 }';
var obj = JSON.parse(str);

Parsing a file containing JSON data

You'll have to do some file operations with fs module.

Asynchronous version

var fs = require('fs');

fs.readFile('/path/to/file.json', 'utf8', function (err, data) {
    if (err) throw err; // we'll not consider error handling for now
    var obj = JSON.parse(data);
});

Synchronous version

var fs = require('fs');
var json = JSON.parse(fs.readFileSync('/path/to/file.json', 'utf8'));

You wanna use require? Think again!

You can sometimes use require:

var obj = require('path/to/file.json');

But, I do not recommend this for several reasons:

1. require is synchronous. If you have a very big JSON file, it will choke your event loop. You really need to use JSON.parse with fs.readFile.
2. require will read the file only once. Subsequent calls to require for the same file will return a cached copy. Not a good idea if you want to read a .json file that is continuously updated. You could use a hack. But at this point, it's easier to simply use fs.
3. If your file does not have a .json extension, require will not treat the contents of the file as JSON.

Seriously! Use JSON.parse.

load-json-file module

If you are reading large number of .json files, (and if you are extremely lazy), it becomes annoying to write boilerplate code every time. You can save some characters by using the load-json-file module.

const loadJsonFile = require('load-json-file');

Asynchronous version

loadJsonFile('/path/to/file.json').then(json => {
    // `json` contains the parsed object
});

Synchronous version

let obj = loadJsonFile.sync('/path/to/file.json');

Parsing JSON from streams

If the JSON content is streamed over the network, you need to use a streaming JSON parser. Otherwise it will tie up your processor and choke your event loop until JSON content is fully streamed.

There are plenty of packages available in NPM for this. Choose what's best for you.

Error Handling/Security

If you are unsure if whatever that is passed to JSON.parse() is valid JSON, make sure to enclose the call to JSON.parse() inside a try/catch block. A user provided JSON string could crash your application, and could even lead to security holes. Make sure error handling is done if you parse externally-provided JSON.

If you want to add some comments in your JSON and allow trailing commas you might want use below implemention:

var fs = require('fs');

var data = parseJsData('./message.json');

console.log('[INFO] data:', data);

function parseJsData(filename) {
    var json = fs.readFileSync(filename, 'utf8')
        .replace(/\s*\/\/.+/g, '')
        .replace(/,(\s*\})/g, '}')
    ;
    return JSON.parse(json);
}

Note that it might not work well if you have something like "abc": "foo // bar" in your JSON. So YMMV.

Another example of JSON.parse :

var fs = require('fs');
var file = __dirname + '/config.json';

fs.readFile(file, 'utf8', function (err, data) {
  if (err) {
    console.log('Error: ' + err);
    return;
  }

  data = JSON.parse(data);

  console.dir(data);
});

I'd like to mention that there are alternatives to the global JSON object. JSON.parse and JSON.stringify are both synchronous, so if you want to deal with big objects you might want to check out some of the asynchronous JSON modules.

Have a look: https://github.com/joyent/node/wiki/Modules#wiki-parsers-json

Use JSON.parse(str);. Read more about it here.

Here are some examples:

var jsonStr = '{"result":true, "count":42}';

obj = JSON.parse(jsonStr);

console.log(obj.count);    // expected output: 42
console.log(obj.result);   // expected output: true

You can use JSON.parse() (which is a built in function that will probably force you to wrap it with try-catch statements).

Or use some JSON parsing npm library, something like json-parse-or

Parsing a JSON stream? Use JSONStream.

var request = require('request')
  , JSONStream = require('JSONStream')

request({url: 'http://isaacs.couchone.com/registry/_all_docs'})
    .pipe(JSONStream.parse('rows.*'))
    .pipe(es.mapSync(function (data) {
      return data
    }))

https://github.com/dominictarr/JSONStream