[node.js 8+] Recursively get all files in a directory

getFiles.js

// Node 8+
// --------------------------------------------------------------
// No external dependencies

const { promisify } = require('util');
const { resolve } = require('path');
const fs = require('fs');
const readdir = promisify(fs.readdir);
const stat = promisify(fs.stat);

async function getFiles(dir) {
  const subdirs = await readdir(dir);
  const files = await Promise.all(subdirs.map(async (subdir) => {
    const res = resolve(dir, subdir);
    return (await stat(res)).isDirectory() ? getFiles(res) : res;
  }));
  return files.reduce((a, f) => a.concat(f), []);
}

// Usage

getFiles(__dirname)
  .then(files => console.log(files))
  .catch(e => console.error(e));


// Node 10.10+
// --------------------------------------------------------------
// Updated for node 10+ with even more whizbang:

const { resolve } = require('path');
const { readdir } = require('fs').promises;

async function getFiles(dir) {
  const dirents = await readdir(dir, { withFileTypes: true });
  const files = await Promise.all(dirents.map((dirent) => {
    const res = resolve(dir, dirent.name);
    return dirent.isDirectory() ? getFiles(res) : res;
  }));
  return Array.prototype.concat(...files);
}

// Note that starting with node 11.15.0 you can use 
// `files.flat()` instead of `Array.prototype.concat(...files)` 
// to flatten the array.


// Node 11+
// --------------------------------------------------------------
// The following version uses async iterators.

const { resolve } = require('path');
const { readdir } = require('fs').promises;

async function* getFiles(dir) {
  const dirents = await readdir(dir, { withFileTypes: true });
  for (const dirent of dirents) {
    const res = resolve(dir, dirent.name);
    if (dirent.isDirectory()) {
      yield* getFiles(res);
    } else {
      yield res;
    }
  }
}

// Usage has changed because the return type is now an 
// async iterator instead of a promise:

(async () => {
  for await (const f of getFiles('.')) {
    console.log(f);
  }
})()

// I've written more about async iterators here: 
// https://qwtel.com/posts/software/async-generators-in-the-wild/

Others please correct me if I’m wrong, following is based on my recent understanding of promises and async/await world 😅

As the readdir fonction is declared and assigned as the promisified version of the async/node style callback fs.readDir function, second line which calls it does not block thread execution, under the hood it is still an async op for reading entry directory

await keyword before a call to a function which returns a promise (or defined as an async function) allow to wait for and get the result returned by the operation executed by the promise itself, and provide a way for coder to have a simpler code syntax and code style familiar and resembling the traditional synchronous way of writing code. But other operations in the background of this code may be executed while «awaiting » for the result

In the above function, line 3 gets executed only after the complete execution of line 2 (i.e. when readdir returns ) right ?

Right.

If yes, then the thread is blocked isn't it ?

No. When it hits line 2, it sees the await and does 2 things:

• it creates a function.then(<after await code here>) scope
• continues executing the next instruction in the event loop

await is an easier version of saying myFunction().then(anotherFunction)

So the desugared version of this is:

async function getFiles(dir) {
	const subdirs = await readdir(dir);
	/* after await code here */
}

// desugared

function getFiles(dir) {
	return readdir(dir).then(function(){
		/* after await code here */
	})
}

As you can see, with the desugared version, there is no blocking whatsoever.

Thanks mrchief.
I understood how it works now.
By the way, where did you get this information from?
It is very hard to surf for this explaintion...

Thanks again.... :)

There's a good amount of explanation in MDN docs. And if you google for it, you can find many online resources covering the same topic as well.

For example "Node 10.10+":

// Node 10.10+
// --------------------------------------------------------------
// Updated for node 10+ with even more whizbang:

const { resolve } = require('path');
const { readdir } = require('fs').promises;

async function getFiles(dir) {
  const dirents = await readdir(dir, { withFileTypes: true });
  const files = dirents.map((dirent) => {
    const res = resolve(dir, dirent.name);
    return dirent.isDirectory() ? getFiles(res) : res;
  });
  return Array.prototype.concat(...files);
}

That doesn't work properly because it returns unresolved promises. This should work:

async function getFiles(dir) {
  const dirents = await readdir(dir, { withFileTypes: true });

  const files = await Promise.all(dirents.map((dirent) => {
    const res = resolve(dir, dirent.name);
    return dirent.isDirectory() ? getFiles(res) : res;
  }));

  return Array.prototype.concat(...files);
}

That doesn't work properly because it returns unresolved promises. This should work:

What an oversight. I've fixed it in the gist. Thanks for posting a correction.

@qwtel
Great article/examples. Despite having years of embedded multitasking OS development background, Promises just have so many gotchas and ambiguities from my perspective. Still anyway question:

I'm thoroughly confused about the following statements:

const res = resolve(dir, subdir)

Everything I have read is very clear that Promise.resolve() can only return one value. Using any object/array can obviously get around that.
What paradigm allows the above resolve with two parameters? More importantly resolve is undefined (executing under Thunderbird)
with a directory iterator so I had to do some adaptation, but I don't think the resolve question changes.
Thanks
@cleidigh

@cleidigh resolve here is actually path.resolve which takes multiple paths and returns a single value.

Note the const { resolve } = require('path'); in the snippet.

@mrchief
Thanks! I got completely hung up on the resolve and missed the reference to the path module.
I'm trying to adapt this using a Mozilla Thunderbird DirectoryInterator as opposed to readDir.
I have been trying every possible combination of promise handling , but I am obviously flummoxed.
Removing the unnecessary path combination in my case still fails.
I'm going to keep experimenting but I may post my best if I can't get it. This approach is certainly what I need to use just with some tweaks that Confound me.