52

The crypto module of node.js (at the time of this writing at least) is not still deemed stable and so the API may change. In fact, the methods that everyone in the internet use to get the hash (md5, sha1, ...) of a file are considered legacy (from the documentation of Hash class) (note: emphasis mine):

Class: Hash

The class for creating hash digests of data.

It is a stream that is both readable and writable. The written data is used to compute the hash. Once the writable side of the stream is ended, use the read() method to get the computed hash digest. The legacy update and digest methods are also supported.

Returned by crypto.createHash.

Despite hash.update and hash.digest being considered legacy, the example shown just above the quoted snippet is using them.

What's the correct way of obtaining hashes without using those legacy methods?

Carlos Campderrós
  • 22,354
  • 11
  • 51
  • 57

6 Answers6

96

From the quoted snippet in the question:

[the Hash class] It is a stream that is both readable and writable. The written data is used to compute the hash. Once the writable side of the stream is ended, use the read() method to get the computed hash digest.

So what you need to hash some text is:

var crypto = require('crypto');

// change to 'md5' if you want an MD5 hash
var hash = crypto.createHash('sha1');

// change to 'binary' if you want a binary hash.
hash.setEncoding('hex');

// the text that you want to hash
hash.write('hello world');

// very important! You cannot read from the stream until you have called end()
hash.end();

// and now you get the resulting hash
var sha1sum = hash.read();

If you want to get the hash of a file, the best way is create a ReadStream from the file and pipe it into the hash:

var fs = require('fs');
var crypto = require('crypto');

// the file you want to get the hash    
var fd = fs.createReadStream('/some/file/name.txt');
var hash = crypto.createHash('sha1');
hash.setEncoding('hex');

fd.on('end', function() {
    hash.end();
    console.log(hash.read()); // the desired sha1sum
});

// read all file and pipe it (write it) to the hash object
fd.pipe(hash);
geon
  • 8,128
  • 3
  • 34
  • 41
Carlos Campderrós
  • 22,354
  • 11
  • 51
  • 57
  • is it possible to get the data content from hash object? – sunnycmf Feb 19 '14 at 18:40
  • @sunnycmf what data do you mean? If you mean the original data you were hashing then I don't think so. If you mean the computed hash, then just use `hash.read()`. – Carlos Campderrós Feb 20 '14 at 08:47
  • yes i mean orig data, coz want to read the file once and get the data & sha1 hash. – sunnycmf Feb 21 '14 at 16:49
  • @sunnycmf you can use pipe twice: file.pipe(hash) and then file.pipe(outputStream) – Jacopofar Dec 01 '15 at 16:39
  • 7
    Instead of the `fd.on('end')`, it would be better to handle `hash.on('finish')`, which does not require calling `hash.end()`. – Herman Kan May 06 '16 at 10:03
  • 1
    Seems to me the pipe-to-Hash-object method is much more fickle than the old-fashioned `stream.on('data', data => hash.update(data))` and later calling `digest('hex')`. When I did the exact experiment as above but left out `hash.read()` because I wanted to do that on the `finish` event of the write-stream I got nothing, no value. Only when I call `hash.read()` in the `end` event handler as above (or in `finish` suggested by @HermanKan) do I get the hash. No such problems with calling `hash.update(...)` in a `data` event handler, then I can get the hash output anywhere, any time. – Mörre Jul 07 '16 at 12:30
44

An ES6 version returning a Promise for the hash digest:

function checksumFile(hashName, path) {
  return new Promise((resolve, reject) => {
    const hash = crypto.createHash(hashName);
    const stream = fs.createReadStream(path);
    stream.on('error', err => reject(err));
    stream.on('data', chunk => hash.update(chunk));
    stream.on('end', () => resolve(hash.digest('hex')));
  });
}
Community
  • 1
  • 1
DS.
  • 22,632
  • 6
  • 47
  • 54
  • 2
    I like this. Possible improvement here: https://gist.github.com/F1LT3R/2e4347a6609c3d0105afce68cd101561 – f1lt3r Aug 01 '17 at 04:10
21

Short version of Carlos' answer:

var fs = require('fs')
var crypto = require('crypto')

fs.createReadStream('/some/file/name.txt').
  pipe(crypto.createHash('sha1').setEncoding('hex')).
  on('finish', function () {
    console.log(this.read()) //the hash
  })
Afanasii Kurakin
  • 3,330
  • 2
  • 24
  • 26
16

Further polish, ECMAScript 2015

hash.js:

'use strict';

function checksumFile(algorithm, path) {
  return new Promise(function (resolve, reject) {
    let fs = require('fs');
    let crypto = require('crypto');

    let hash = crypto.createHash(algorithm).setEncoding('hex');
    fs.createReadStream(path)
      .once('error', reject)
      .pipe(hash)
      .once('end', function () {
        resolve(hash.read());
      });
  });
}

checksumFile('sha1', process.argv[2]).then(function (hash) {
  console.log('hash:', hash);
});

node hash.js hash.js
hash: 9c92ec7acf75f943aac66ca17427a4f038b059da

Works at least as early as v10.x:

node --version
v10.24.1
Bohdan Yurov
  • 365
  • 5
  • 9
Harald Rudell
  • 787
  • 6
  • 7
6

I use Node module hasha successfully, the code becomes very clean and short. It returns a promise, so you can use it with await:

const hasha = require('hasha');

const fileHash = await hasha.fromFile(yourFilePath, {algorithm: 'md5'});
Lacoste
  • 89
  • 2
  • 6
0
var fs = require('fs');
var crypto = require('crypto');
var fd = fs.createReadStream('data.txt');
var hash = crypto.createHash('md5');
hash.setEncoding('hex');
fd.pipe(hash);
hash.on('data', function (data) {
    console.log('# ',data);
});
javelinthrow
  • 9
  • 1