784

I'm developing a console script for personal needs. I need to be able to pause for an extended amount of time, but, from my research, Node.js has no way to stop as required. It’s getting hard to read users’ information after a period of time... I’ve seen some code out there, but I believe they have to have other code inside of them for them to work such as:

    setTimeout(function() {
    }, 3000);

However, I need everything after this line of code to execute after the period of time.

For example,

    // start of code
    console.log('Welcome to my console,');

    some-wait-code-here-for-ten-seconds...

    console.log('Blah blah blah blah extra-blah');
    // end of code

I've also seen things like

    yield sleep(2000);

But Node.js doesn't recognize this.

How can I achieve this extended pause?

4
  • 4
    @Christopher Allen, Maybe not relevant, but does the job: require("child_process").execSync('php -r "sleep($argv[1]);" ' + seconds); Commented Jan 4, 2016 at 9:57
  • The node-sleep npm module might do the trick (however, I would only use it for debugging) Commented Feb 19, 2016 at 2:22
  • 1
    Does this answer your question? What is the JavaScript version of sleep()? Commented May 24, 2021 at 8:04
  • 2
    Please don't write your own promises! Use import { setTimeout } from 'timers/promises'. Commented Feb 22, 2023 at 13:51

27 Answers 27

959

Update Jan 2021: You can even do it in the Node REPL interactive using --experimental-repl-await flag

$ node --experimental-repl-await
> const delay = ms => new Promise(resolve => setTimeout(resolve, ms))
> await delay(1000) /// waiting 1 second.

A new answer to an old question. Today ( Jan 2017 June 2019) it is much easier. You can use the new async/await syntax. For example:

async function init() {
  console.log(1);
  await sleep(1000);
  console.log(2);
}

function sleep(ms) {
  return new Promise((resolve) => {
    setTimeout(resolve, ms);
  });
}

For using async/await out of the box without installing and plugins, you have to use node-v7 or node-v8, using the --harmony flag.

Update June 2019: By using the latest versions of NodeJS you can use it out of the box. No need to provide command line arguments. Even Google Chrome support it today.

Update May 2020: Soon you will be able to use the await syntax outside of an async function. In the top level like in this example

await sleep(1000)
function sleep(ms) {
  return new Promise((resolve) => {
    setTimeout(resolve, ms);
  });
}

The proposal is in stage 3. You can use it today by using webpack 5 (alpha),

Before that becomes available, you can just wrap the script toplevel in a self calling async function:

(async function() {
const sleep = ms => new Promise(resolve => setTimeout(resolve, ms))
console.log(1)
await sleep(1000)
console.log(2)
})()

More info:

9
  • 5
    I think you forgot the await keyword in front of sleep(1000) Commented Mar 13, 2017 at 17:56
  • 7
    This really should be the answer, the sleep node.js package can be troublesome. Commented Dec 20, 2017 at 15:16
  • 69
    let sleep = ms => new Promise(resolve => setTimeout(resolve, ms)); For oneliner freaks like myself :) Commented May 30, 2018 at 8:25
  • 30
    let sleep = require('util').promisify(setTimeout); works on Node 7.6+ and improves readability
    – Brian H.
    Commented May 6, 2019 at 21:29
  • 2
    This saved my bacon today (specifically the promisify recipe @BrianHVB posted).
    – kindall
    Commented May 15, 2019 at 20:11
777

The shortest solution without any dependencies:

await new Promise(resolve => setTimeout(resolve, 5000));
12
  • 74
    "The height of sophistication is simplicity." -- Clare Booth Luce. This is by far the best answer, IMO.
    – arnold
    Commented Jun 28, 2018 at 20:30
  • 3
    This is obviously much more recent than the other answers, but it's the most elegant as of 2018 that gets the job done in 1 line without any other impacts to the code.
    – Herick
    Commented Dec 3, 2018 at 19:39
  • 1
    This is a good one. You have to be using NodeJS 7.6.0 or above though. It won't work on older versions. Commented Dec 30, 2018 at 22:06
  • 10
    let sleep = require('util').promisify(setTimeout); is three characters longer but reusable and more readable imo
    – Brian H.
    Commented May 6, 2019 at 21:32
  • 2
    To avoid an eslint complaint, call it resolve instead of done. I.e. await new Promise(resolve => setTimeout(resolve, 5000)) Commented Nov 2, 2019 at 18:40
253

Best way to do this is to break your code into multiple functions, like this:

function function1() {
    // stuff you want to happen right away
    console.log('Welcome to My Console,');
}

function function2() {
    // all the stuff you want to happen after that pause
    console.log('Blah blah blah blah extra-blah');
}

// call the first chunk of code right away
function1();

// call the rest of the code and have it execute after 3 seconds
setTimeout(function2, 3000);

It's similar to JohnnyHK's solution, but much neater and easier to extend.

5
  • 8
    @LucasSeveryn You're doing something wrong, then. This is a core design pattern. Commented Jun 9, 2015 at 17:23
  • And what do you do when the start of the function is not just one function away, but 10 files away from the code that needs to be de-asynchronized ? Commented Jun 28, 2016 at 14:14
  • 22
    @CyrilDuchon-Doris What?
    – AturSams
    Commented May 3, 2017 at 15:39
  • 3
    use @machineghost's answer for an elegant promise based solution that requires no custom code and supports await/async Commented Nov 27, 2017 at 19:50
  • This is async, it means that if function1 terminates before 3 seconds, they start overlapping each other. Then you can't even return and it's almost never utilizable. The only way I found so far was using debugger; Commented Aug 16, 2018 at 23:14
204

This is a simple blocking technique:

var waitTill = new Date(new Date().getTime() + seconds * 1000);
while(waitTill > new Date()){}

It's blocking insofar as nothing else will happen in your script (like callbacks). But since this is a console script, maybe it is what you need!

12
  • 44
    horrible or not, it does a simple wait, it blocks and it works for some testing purpose. Exactly what I was searching for.
    – Clijsters
    Commented Jul 10, 2016 at 20:48
  • 15
    perfectly answers the question without 3rd party libs, and is simple. Yet people say "horrible" .... This is great e.g. for simulating heavy CPU load etc. Btw very similar to this phpied.com/sleep-in-javascript Commented Feb 6, 2017 at 17:24
  • 27
    This is a spin-wait.
    – Mike Atlas
    Commented Mar 29, 2017 at 16:14
  • 12
    @Ali that seems to be the goal. Commented Mar 21, 2018 at 19:22
  • 8
    This is the only correct answer to the question. The OP asked specifically to call a wait function, return from the function, and continue. (No callbacks, no async anything, no rewriting your whole call stack to handle promises.) The only other solution that does exactly what the OP wanted is to call an external program, for example SLEEP.EXE on Windows.
    – blitter
    Commented Apr 22, 2020 at 16:22
167

Put the code that you want executed after the delay within the setTimeout callback:

console.log('Welcome to My Console,');
setTimeout(function() {
    console.log('Blah blah blah blah extra-blah');
}, 3000);
5
  • 5
    This is terribly messy and generally bad practice, especially if the OP wants the rest of the program to run after that delay. See my answer. Commented Jan 10, 2013 at 1:48
  • 38
    @ElliotBonneville It's just an example to illustrate the concept. Obviously you could (should) factor the code into a function call instead instead of using inline code, just like anywhere else.
    – JohnnyHK
    Commented Jan 10, 2013 at 1:51
  • @ChristopherKemp: Turns out Node.js has a solution for this called node-fibers. Check it out. Commented Jan 10, 2013 at 1:55
  • This is a great solution, particularly if you don't want to execute any code, you just want to mimic a "sleep" method inside a loop. Nothing wrong with this example.
    – Halfstop
    Commented Oct 25, 2016 at 17:56
  • this is perfect for delaying on requests coming from client, I used this approach to test my loading spinners on client side Commented Dec 24, 2017 at 6:54
122

On Node 7.6.0 or higher

Node supports waiting natively:

const sleep = (waitTimeInMs) => new Promise(resolve => setTimeout(resolve, waitTimeInMs));

then if you can use async functions:

await sleep(10000); // sleep for 10 seconds

or:

sleep(10000).then(() => {
  // This will execute 10 seconds from now
});

On older Node versions (original answer)

I wanted an asynchronous sleep that worked in Windows & Linux, without hogging my CPU with a long while loop. I tried the sleep package but it wouldn't install on my Windows box. I ended up using:

https://www.npmjs.com/package/system-sleep

To install it, type:

npm install system-sleep

In your code,

var sleep = require('system-sleep');
sleep(10*1000); // sleep for 10 seconds

Works like a charm.

13
  • 3
    great answer thanks - it made some impossible code possible - this is NOT a spin-wait
    – danday74
    Commented Sep 9, 2017 at 23:55
  • 1
    On further research this module relies on deasync which is forked from another deasync github repo. The original repo warns not to use it as it is a hack. It does work but not on all platforms so if you need a cross platform solution avoid this one.
    – danday74
    Commented Sep 11, 2017 at 7:17
  • 1
    yes, i've never experienced probs with it and its great for dev - under the hood I believe it relies on C or C++ which is widely available but I guess on some machines its not and thats when it fails which is why it causes issues - if it fails system sleep falls back to a spin wait which will freeze code execution
    – danday74
    Commented Mar 8, 2018 at 7:16
  • 2
    This package doesn't work on Mac OS, and is therefore not cross compatible, and therefore not workable. See github.com/jochemstoel/nodejs-system-sleep/issues/4
    – marknuzz
    Commented Dec 29, 2018 at 14:04
  • 2
    This should be higher. Brilliant answer
    – dz210
    Commented Jun 4, 2021 at 7:47
76

From Node.js 15 and up you can use the Timers Promises API. You don't have to promisify setTimeout or rely on a 3rd party library anymore.

import { setTimeout } from 'node:timers/promises';

await setTimeout(1000);
2
  • 5
    This is the correct solution - please don't write your own promises. Commented Feb 22, 2023 at 13:49
  • 1
    Absolutely the correct solution. Commented Sep 13, 2023 at 15:06
70

Simple and elegant sleep function using modern Javascript

function sleep(millis) {
    return new Promise(resolve => setTimeout(resolve, millis));
}

No dependencies, no callback hell; that's it :-)


Considering the example given in the question, this is how we would sleep between two console logs:

async function main() {
    console.log("Foo");
    await sleep(2000);
    console.log("Bar");
}

main();

The "drawback" is that your main function now has to be async as well. But, considering you are already writing modern Javascript code, you are probably (or at least should be!) using async/await all over your code, so this is really not an issue. All modern browsers today support it.

Giving a little insight into the sleep function for those that are not used to async/await and fat arrow operators, this is the verbose way of writing it:

function sleep(millis) {
    return new Promise(function (resolve, reject) {
        setTimeout(function () { resolve(); }, millis);
    });
}

Using the fat arrow operator, though, makes it even smaller (and more elegant).

3
  • 1
    You can also write the sleep function without the async and leaving everything else the same. It is probably clearer with the async though. Commented Oct 27, 2017 at 5:55
  • Good catch, @KevinPeña. Matter of fact, I think I prefer it without async. Edited my answer with your suggestion. I don't think it makes the code clearer; for that, I'd resort to JSDoc instead. Commented Oct 28, 2017 at 22:46
  • 9
    I like to have the following one-liner in my scripts: const sleep = ms => new Promise(resolve => setTimeout(resolve, ms));
    – korya
    Commented Jan 23, 2018 at 18:40
53

You can use this www.npmjs.com/package/sleep

var sleep = require('sleep');
sleep.sleep(10); // sleep for ten seconds
3
  • 5
    It works fine on MacOS, but it encounter errors on CentOS due to node_gyp errors. It seems not portable.
    – AechoLiu
    Commented Aug 11, 2017 at 7:30
  • 2
    perhaps problem caused not by OS, but by node build
    – Urgotto
    Commented Oct 7, 2017 at 9:23
  • This is a newer feature, and so will require a newer version of Node. Commented Dec 18, 2021 at 22:41
44

If you want to "code golf" you can make a shorter version of some of the other answers here:

const sleep = ms => new Promise(resolve => setTimeout(resolve, ms));

But really the ideal answer in my opinion is to use Node's util library and its promisify function, which is designed for exactly this sort of thing (making promise-based versions of previously existing non-promise-based stuff):

const util = require('util');
const sleep = util.promisify(setTimeout);

In either case you can then pause simply by using await to call your sleep function:

await sleep(1000); // sleep for 1s/1000ms

EDIT: As noted in the comments, you can even reduce that to one line:

const sleep = require('util').promisify(setTimeout);

Or, if you don't even want to bother making a sleep function:

await require('util').promisify(setTimeout)(1000);
4
  • Actually, using util.promisify, it is not possible to call sleep without also providing a callback. nodejs.org/api/…
    – Howie
    Commented Dec 17, 2017 at 22:19
  • 1
    You're missing the await. It sort of creates a callback, pauses things, and then restarts the code when that callback returns. The callback returns with a value, but we don't care about it in this case, so there's nothing to the left of await. Commented Dec 18, 2017 at 1:29
  • 1
    write it in one line to wine the code golf ;) const sleep = require('util').promisify(setTimeout);
    – Fabian
    Commented Jan 4, 2019 at 15:05
  • 2
    @Fabian, and if you need it only once: await require('util').promisify(setTimeout)(1000);
    – Ed'ka
    Commented Sep 26, 2020 at 0:41
35

This question is quite old, but recently V8 has added Generators which can accomplish what the OP requested. Generators are generally easiest to use for async interactions with the assistance of a library such as suspend or gen-run.

Here's an example using suspend:

suspend(function* () {
    console.log('Welcome to My Console,');
    yield setTimeout(suspend.resume(), 10000); // 10 seconds pass..
    console.log('Blah blah blah blah extra-blah');
})();

Related reading (by way of shameless self promotion): What's the Big Deal with Generators?.

5
  • 2
    Good answer - But should read yield setTimeout(suspend.resume(), 10000);
    – edhubbell
    Commented Apr 22, 2015 at 18:19
  • 1
    Thanks, @edhubbell. This answer was based on a very old version of suspend, but you're right regarding the latest. I'll update the answer.
    – jmar777
    Commented Apr 23, 2015 at 5:32
  • what version of node is this for?
    – danday74
    Commented Sep 9, 2017 at 21:28
  • 1
    @danday74 I can't recall exactly when they were un-flagged, but they've been around since v0.12 behind the --harmony flag, and according to node.green, they're at least available without any flags since v4.8.4: node.green/#ES2015-functions-generators-basic-functionality. Please note, however, that the newer async/await syntax provides a better solution to this now, with no need for extra libraries like suspend. See this answer for an example: stackoverflow.com/a/41957152/376789. Async functions are available (without flags) since v7.10.
    – jmar777
    Commented Sep 12, 2017 at 15:28
  • This is very complicated as of 2021. Commented May 24, 2021 at 8:05
30

Try using promise, it works for me in NodeJS

one liner

await new Promise(resolve => setTimeout(resolve, 5000));

or have it as a function in NodeJS to re-use

const sleep = async (milliseconds) => {
    await new Promise(resolve => setTimeout(resolve, milliseconds));
}

use the function like

await sleep(5000)
23

On Linux/nodejs this works for me:

const spawnSync = require('child_process').spawnSync;

var sleep = spawnSync('sleep', [1.5]);

It is blocking, but it is not a busy wait loop.

The time you specify is in seconds but can be a fraction. I don't know if other OS's have a similar command.

1
  • sleep is pretty much omnipresent and a utiltiy arround since early UNIX days en.wikipedia.org/wiki/Sleep_%28Unix%29. Only "not rare" os it would maybe not exist is windows (however there you could attempt child_process.spawnSync('timeout', ['/T', '10']) Commented Oct 24, 2018 at 16:41
22

Node 16 has a new way to do it easily

import { setTimeout } from 'timers/promises'

console.log('before')
await setTimeout(3000)
console.log('after')
1
14

I've recently created simpler abstraction called wait.for to call async functions in sync mode (based on node-fibers). There is also a version based on upcoming ES6 Generators.

https://github.com/luciotato/waitfor

Using wait.for, you can call any standard nodejs async function, as if it were a sync function, without blocking node's event loop.

You can code sequentially when you need it, which is, (I'm guessing) perfect to simplify your scripts for personal use.

using wait.for your code will be:

require('waitfor')

..in a fiber..
//start-of-code
console.log('Welcome to My Console,');
wait.miliseconds(10*1000); //defined in waitfor/paralell-tests.js - DOES NOT BLOCK
console.log('Blah blah blah blah extra-blah');
//endcode. 

Also any async function can be called in Sync mode. Check the examples.

4
  • TypeError: Object #<Object> has no method 'miliseconds' Commented Mar 8, 2014 at 23:59
  • the comment says: "//defined in waitfor/paralell-tests.js" grab it from that file. Commented Mar 14, 2014 at 5:59
  • 2
    after I got it from wait.for/paralell-tests.js I encountered a another errors related to undefined properties etc. So I needed to copy them too. Why don't you organize the code in a way that this will not be required?
    – d.k
    Commented Mar 27, 2014 at 0:06
  • Wait.for and other fiber solutions have opened up a whole new world to me! I'd up vote this a million times if I could. Although most of the nodejs community oppose fibers, I think they're a fantastic addition and definitely have their place when it comes to callback hell. Commented May 23, 2014 at 21:09
10

Since, javascript engine (v8) runs code based on sequence of events in event-queue, There is no strict that javascript exactly trigger the execution at after specified time. That is, when you set some seconds to execute the code later, triggering code is purely base on sequence in event queue. So triggering execution of code may take more than specified time.

So Node.js follows,

process.nextTick()

to run the code later instead setTimeout(). For example,

process.nextTick(function(){
    console.log("This will be printed later");
});
3

With ES6 supporting Promises, we can use them without any third-party aid.

const sleep = (seconds) => {
    return new Promise((resolve, reject) => {
        setTimeout(resolve, (seconds * 1000));
    });
};

// We are not using `reject` anywhere, but it is good to
// stick to standard signature.

Then use it like this:

const waitThenDo(howLong, doWhat) => {
    return sleep(howLong).then(doWhat);
};

Note that the doWhat function becomes the resolve callback within the new Promise(...).

Also note that this is ASYNCHRONOUS sleep. It does not block the event loop. If you need blocking sleep, use this library which realizes blocking sleep with the help of C++ bindings. (Although the need for a blocking sleep in Node like async environments is rare.)

https://github.com/erikdubbelboer/node-sleep

1

In order to "wait" in javascript using promises are the way to go as the top answers show.

So how can it be used?

Here's a simple example of a 5-second sub-process queuing up parameters for a 4-second main process in a non-blocking manner.

const wait = (seconds) => 
    new Promise(resolve => 
        setTimeout(() => 
            resolve(true), seconds * 1000))

const process = async (items, prepTask, mainTask) => {
    const queue = [];
    let done = false;

    items.forEach((item, i) => {
        prepTask(item).then(() => {
            queue.push(item);
            if (i == items.length -1) {
                done = true;
            }
        })
    })

    while (!done || queue.length) {
        if (queue.length) {
            const workload = queue.shift();
            await mainTask(workload)
        } else {
            console.log('waiting for subtask to queue')
            await wait(1);
        }
    }
}

// Usage Example

const ids = [1,2,3,4,5,6,7,8,9,10];

const prepTask = async (id) => {
    await wait(id * 5)
    return id * 5;
}

const mainTask = async (workload) => {
    console.log('excuting workload: ', workload);
    const result = await wait(4);
    return { workload, result }
}

process(ids, prepTask, mainTask)
    .then(() => console.log('done'))
0
let co = require('co');
const sleep = ms => new Promise(res => setTimeout(res, ms));

co(function*() {
    console.log('Welcome to My Console,');
    yield sleep(3000);
    console.log('Blah blah blah blah extra-blah');
});

This code above is the side effect of the solving Javascript's asynchronous callback hell problem. This is also the reason I think that makes Javascript a useful language in the backend. Actually this is the most exciting improvement introduced to modern Javascript in my opinion. To fully understand how it works, how generator works needs to be fully understood. The function keyword followed by a * is called a generator function in modern Javascript. The npm package co provided a runner function to run a generator.

Essentially generator function provided a way to pause the execution of a function with yield keyword, at the same time, yield in a generator function made it possible to exchange information between inside the generator and the caller. This provided a mechanism for the caller to extract data from a promise from an asynchronous call and to pass the resolved data back to the generator. Effectively, it makes an asynchronous call synchronous.

2
  • While this code may answer the question, providing additional context regarding how and/or why it solves the problem would improve the answer's long-term value. Commented Mar 8, 2017 at 11:17
  • Thanks @DonaldDuck. The code in my answer is probably the most exciting part of improvement in Javascript. I'm so amazed that some super smart people think about this way to solve the callback hell problem.
    – user1663023
    Commented Mar 8, 2017 at 12:03
0

simple we are going to wait for 5 seconds for some event to happen (that would be indicated by done variable set to true somewhere else in the code) or when timeout expires that we will check every 100ms

    var timeout=5000; //will wait for 5 seconds or untildone
    var scope = this; //bind this to scope variable
    (function() {
        if (timeout<=0 || scope.done) //timeout expired or done
        {
            scope.callback();//some function to call after we are done
        }
        else
        {
            setTimeout(arguments.callee,100) //call itself again until done
            timeout -= 100;
        }
    })();
0

For some people, the accepted answer is not working, I found this other answer and it is working for me: How can I pass a parameter to a setTimeout() callback?

var hello = "Hello World";
setTimeout(alert, 1000, hello); 

'hello' is the parameter being passed, you can pass all the parameters after the timeout time. Thanks to @Fabio Phms for the answer.

0

This is a moment.js flavored module based on the dirty blocking approach suggested by @atlex2. Use this only for testing.

const moment = require('moment');

let sleep = (secondsToSleep = 1) => {
    let sleepUntill = moment().add(secondsToSleep, 'seconds');
    while(moment().isBefore(sleepUntill)) { /* block the process */ }
}

module.exports = sleep;
0
function doThen(conditional,then,timer) {
    var timer = timer || 1;
    var interval = setInterval(function(){
        if(conditional()) {
            clearInterval(interval);
            then();
        }
    }, timer);
}

Example usage:

var counter = 1;
doThen(
    function() {
        counter++;
        return counter == 1000;
    },
    function() {
        console.log("Counter hit 1000"); // 1000 repeats later
    }
)
-1

If you just need to suspend for testing purpose you current thread execution try this:

function longExecFunc(callback, count) {

    for (var j = 0; j < count; j++) {
        for (var i = 1; i < (1 << 30); i++) {
            var q = Math.sqrt(1 << 30);
        }
    }
    callback();
}
longExecFunc(() => { console.log('done!')}, 5); //5, 6 ... whatever. Higher -- longer
-2

The other answers are great but I thought I'd take a different tact.

If all you are really looking for is to slow down a specific file in linux:

 rm slowfile; mkfifo slowfile; perl -e 'select STDOUT; $| = 1; while(<>) {print $_; sleep(1) if (($ii++ % 5) == 0); }' myfile > slowfile  &

node myprog slowfile

This will sleep 1 sec every five lines. The node program will go as slow as the writer. If it is doing other things they will continue at normal speed.

The mkfifo creates a first-in-first-out pipe. It's what makes this work. The perl line will write as fast as you want. The $|=1 says don't buffer the output.

-3

I put together, after having read the answers in this question, a simple function which can also do a callback, if you need that:

function waitFor(ms, cb) {
  var waitTill = new Date(new Date().getTime() + ms);
  while(waitTill > new Date()){};
  if (cb) {
    cb()
  } else {
   return true
  }
}
-4

For more info on

yield sleep(2000); 

you should check Redux-Saga. But it is specific to your choice of Redux as your model framework (although strictly not necessary).

Not the answer you're looking for? Browse other questions tagged or ask your own question.