I run into this every now and then:
return somethingThatReturnsAPromise()
.then((response) => {
soSomethingg(); // Eg; update the UI
return response;
});
Now I'm looking for something that is not expected to return anything and won't change the promise chain if I forget that:
return somethingThatReturnsAPromise()
.whatImLookingFor((response) => {
doSomething(); // Eg; update the UI
})
.then((response) => {
// and this one should still be able to access response
});
Maybe this goes against the idea of promises, but for me, it's a bit inconvenient since I can't pass arbitrary functions.
One idea is to compose a function:
const sideEffect = (callback) => {
return (response) => {
callback(response);
return response;
};
};
And I could use it as
return somethingThatReturnsAPromise()
.then(sideEffect(doSomething));
But I'd prefer something instead of then is there something like that?
Note: I'm working with Angular 1.x so I need something like for that.
I would assume that you're not really writing .then().then(), because you could collapse that into a single .then, but that your concern is really about returning the promise and having some external code add another then to the chain. In that case do this:
let p = somethingThatReturnsAPromise();
p.then(() => doSomething());
return p;
This allows the caller to attach additional thens to the original promise instead of chaining off of your .then, thereby receiving the original promise's value. This is called branching the promise chain.
Maybe this goes against the idea of promises
Slightly, promise chains are pipelines where then handlers transform things at each stage. But it's perfectly valid to want to pass through the value unchanged.
One idea is to compose a function:
Indeed the first thing that came to mind, and how I'd do it.
But I'd prefer something instead of then is there something like that?
There isn't. You could add it for your own projects (I wouldn't in a library) by adding it to Promise.prototype. Or you could give yourselve a Promise subclass and add it there.
With a Promise sublass you'd do something like:
return MyPromise.resolve(somethingThatReturnsAPromise())
.thenSide(soSomethingg); // Eg; update the UI
...where thenSide is your method that's then but passing the original value back unchanged, e.g.:
class MyPromise extends Promise {
thenSide(callback) {
this.then(callback);
return this;
}
}
or
class MyPromise extends Promise {
thenSide(callback) {
this.then(callback);
return MyPromise.resolve(this);
}
}
...depending on whether you're bothered about thenSide returning the same promise (since then always returns a new one).
As far as I know (I could well be wrong) the wrapper method for "pass-through" side-effects is an idiomatic way to do what you want.
Alternatively (if you need the same response in multiple places) you can break up the promise chain when you encounter a situation like this.
I initially assumed that passing a bare Promise.mapSeries(...) call as an argument to .then() would be the same as wrapping it in a function, like .then(function() { return Promise.mapSeries(...); }). Having written out this question, I'm no longer entirely sure why it works at all.
In the simplified code below, I open a couple of databases asynchronously (openDBAsync()) and then read a file containing a JS object. I use _.map() to iterate over all of the key/value pairs in the object and asynchronously update their values in a database, while keeping track of which ones meet certain criteria (whether the value is odd, in this toy example). Promise.all() waits for all of the async database calls to settle, and then Promise.mapSeries() is used to process each of the subset of keys, which makes another async database call for each one. Finally, I close all the databases.
function processData(path)
{
var oddKeys = [];
return openDBAsync()
.then(function() { return readFileAsync(path); })
.then(function(dataObject) {
return Promise.all(_.map(dataObject, function(value, key) {
if (value % 2) {
oddKeys.push(key);
}
return updateDBAsync(key, ++value);
}))
.then(Promise.mapSeries(
oddKeys,
function(key) {
return updateOddDBAsync(key);
}
))
})
.then(closeDBAsync);
}
The problem is that the database throws errors complaining that I'm trying to update the database after it's been closed. That means that some of the promises generated in the .mapSeries() call are being called after the final closeDBAsync(). I expected all of them to settle before the final .then() call.
If I wrap the call to Promise.mapSeries() in a function:
.then(function() {
return Promise.mapSeries(
oddKeys,
function(key) {
return updateOddDBAsync(key);
}
);
})
Then I don't get any errors. It also works if I put a .delay(2000) before the close database call, which indicates that Promise.mapSeries() isn't settling all of the promises before it finishes.
This seems like either a bug in Bluebird, or, much more likely, I'm not understanding something fundamental about how Promise.mapSeries() works. Any guidance would be much appreciated.
much more likely, I'm not understanding something fundamental about how Promise.mapSeries() works
Nope, this seems to be more a misunderstanding about how .then(…) works.
The then method of promises does always take a callback function (if you're passing anything else [but null], Bluebird should spit out a warning!). By calling .then(Promise.mapSeries(…)), you were passing a promise instead, which is simply ignored. Being ignored, it also is not awaited by anything, which leads to that error of the database being closed too early.
But the direct call to Promise.mapSeries() doesn't get applied to the array immediately. If it was, the array would be empty and this code wouldn't work at all.
Yes it does. Your array is filled from by the _.map callback, which is executed synchronously, before then and mapSeries are invoked.
So the solution is indeed to wrap the call in a function expression, which will only be executed when the Promise.all(…) fulfills, and whose result will then not be ignored but rather awaited. There might be more, different solutions, depending on what degree of parallel execution you want to allow.
Btw, given you are doing a database transaction, your current code is quite fragile. Look into the promise disposer pattern.
I'm facing an troublesome issue, because it does not occurs every time:
I call webservices with angularJS (hosted in Django/Tastypie), with promises sothat WS2 (Web Service #2) is called only when WS1 has been successfully executed, etc.
The code looks like this:
myfunc = function() {
callPromiseWS1()
.then(function(data){
callPromiseWS2()
})
.then(function(data){
callPromiseWS3()
})
};
WS2 is a POST request that writes stuff in Database.
WS3 is a POST request which uses stuff created by WS2.
Sometimes (and not always), WS3 fails because the object that should have been already created by WS2 does NOT exist.
Of course, if I try to execute the WS3 request later manually, it works.
I have the feeling that WS2 and WS3 are executed in parallel, that's not what I expect.
Any idea? Thank you.
You're missing 3 return statements, but the behavior you're experiencing is because of one in particular:
myfunc = function() {
return callPromiseWS1()
.then(function(data){
return callPromiseWS2(); // this one
})
.then(function(data){
return callPromiseWS3();
})
};
If you don't return the promise that callPromiseWS2 returns, the then() won't know that it's dealing with a promise and just move onto the next then as soon as the synchronous code in callPromiseWS2 completes.
Note that if you're not actually using the data parameter in each step (or if you're just going to pass it right into callPromiseWS2 and callPromiseWS3 as the only parameter), you can simplify this and eliminate the need for two of the above return statements and a lot of the code:
myfunc = function() {
return callPromiseWS1()
.then(callPromiseWS2)
.then(callPromiseWS3);
};
Remember that any function that uses promises should either return the promise, or call .done() at the end.
As a matter of personal preference, I think it's a good idea to use a naming scheme for functions that return promises so that you know they return a promise that needs to be dealt with.
I'm not an expert on promises (and maybe the code you posted was simplified), but the first thing I would check is that the first then returns something, or the second .then will have the same object returned by callPromiseWS1() to work on.
I believe it should be something like:
callPromiseWS1()
.then(function(data){
return callPromiseWS2();
})
.then(function(data){
return callPromiseWS3();
})
But, as I said, I'm not an expert on promises and I don't know if it was just a simplified example of your structure.
I'm working on a simple Windows 8 app in which I need to fetch a set of data from a web site. I am using WinJS.xhr() to retrieve this data, which returns a Promise. I then pass a callback into this Promise's .then() method, which supplies my callback with the returned value from the asynchronous call. The .then() method returns another Promise, giving it the value that my callback returns. The basic structure of such a query would be as follows:
WinJS.xhr({ url: "http://www.example.com/" }).then(
function callback( result_from_xhr )
{
//do stuff
return some_value;
}).then(
function secondcallback( some_value )
{
//do stuff
});
In my situation, however, I may need to make additional queries for data depending on the data returned by the first query, and possibly more queries depending on THAT data... and so on, recursively.
I need a way to code this such that the final .then() is not executed until ALL the recursions have completed, similar to this:
function recurse() {
return WinJS.xhr({ url: "http://www.example.com/" }).then(
function callback( result_from_xhr )
{
if( result_from_xhr == something )
{
recurse();
}
});
}
recurse().then(
function final()
{
//finishing code
});
The problem is that, of course, the finishing code is called as soon as the first level of recursion completes. I need some way to nest the new promise and the old promise from within the callback.
I hope my question is clear enough, I'm really not sure how to explain it and frankly the idea of asynchronous recursive code makes my head hurt.
What I would do here is to create a whole new, standalone promise that you can complete manually, and return that from the recurse() function. Then, when you hit the point that you know you're done doing async work, complete that promise.
Promise.join works when you've got a known set of promises - you need the entire array of promises available before you call join. If I followed the original question, you have a variable number of promises, with more possibly popping up as part of async work. Join isn't the right tool in these circumstances.
So, what does this look like? Something like this:
function doSomethingAsync() {
return new WinJS.Promise(function (resolve, reject) {
function recurse() {
WinJS.xhr({ url: "http://www.example.com/" })
.then(function onResult(result_from_xhr) {
if (result_from_xhr === something) {
recurse();
} else {
// Done with processing, trigger the final promise
resolve(whateverValue);
},
function onError(err) {
// Fail everything if one of the requests fails, may not be
// the right thing depending on your requirements
reject(err);
});
}
// Kick off the async work
recurse();
});
}
doSomethingAsync().then(
function final()
{
//finishing code
});
I rearranged where the recursion is happening so that we aren't recreating a new promise every time, thus the nested recurse() function instead of having it at the outer level.
I solved this problem in perhaps a different way (I think it's the same problem) while making my own Windows 8 app.
The reason I came across this problem is because, by default, a query to a table will paginate, and only return 50 results at a time, so I created a pattern to get the first 50, and then the next 50, etc, until a response comes back with less than 50 results, and then resolve the promise.
This code isn't going to be real code, just for illustration:
function getAllRows() {
return new WinJS.Promise(function(resolve, reject){
var rows = [];
var recursivelyGetRows = function(skipRows) {
table.skip(skipRows).read()
.then(function(results){
rows = rows.concat(results);
if (results.length < 50) {
resolve(rows);
} else {
recursivelyGetRows(skipRows + 50);
}
})
}
recursivelyGetRows(0);
});
}
so getAllRows() returns a promise that resolves only after we get a result back with less than 50 results (which indicates it's the last page).
Depending on your use case, you'll probably want to throw an error handler in there too.
In case it's unclear, 'table' is a Mobile Services table.
You will need to use Promise.join().done() pattern. Pass an array of Promises to join(), which in your case would be a collection of xhr calls. The join() will only call done() when all of the xhr Promises have completed. You will get an array of results passed to the done(), which you can then iterate over and start again with a new Promise.join().done() call. The thing to be away of, when using this approach, is that if one of the Promises passed to join() fail, the entire operation is treated as an error condition.
Sorry I don't have time right now to try and stub out the code for you. If I get a chance, I will try to later. But you should be able to insert this into your recursive function and get things to work.
Well, I solved my problem; my recursive function was misinterpreting the data and thus never stopped recursing. Thank you for your help, and I'll be sure to watch those screencasts, as I still don't quite fully grasp the Promise chaining structure.
I’ve been seeing code that looks like:
myObj.doSome("task").then(function(env) {
// logic
});
Where does then() come from?
The traditional way to deal with asynchronous calls in JavaScript has been with callbacks.
Say we had to make three calls to the server, one after the other, to set up our
application. With callbacks, the code might look something like the following (assuming
a xhrGET function to make the server call):
// Fetch some server configuration
xhrGET('/api/server-config', function(config) {
// Fetch the user information, if he's logged in
xhrGET('/api/' + config.USER_END_POINT, function(user) {
// Fetch the items for the user
xhrGET('/api/' + user.id + '/items', function(items) {
// Actually display the items here
});
});
});
In this example, we first fetch the server configuration. Then based on that, we fetch
information about the current user, and then finally get the list of items for the current
user. Each xhrGET call takes a callback function that is executed when the server
responds.
Now of course the more levels of nesting we have, the harder the code is to read, debug,
maintain, upgrade, and basically work with. This is generally known as callback hell.
Also, if we needed to handle errors, we need to possibly pass in another function to each
xhrGET call to tell it what it needs to do in case of an error. If we wanted to have just one
common error handler, that is not possible.
The Promise API was designed to solve this nesting problem and the
problem of error handling.
The Promise API proposes the following:
Each asynchronous task will return a promise object.
Each promise object will have a then function that can take two arguments, a success
handler and an error handler.
The success or the error handler in the then function will be called only once, after
the asynchronous task finishes.
The then function will also return a promise, to allow chaining multiple calls.
Each handler (success or error) can return a value, which will be passed to the next
function as an argument, in the chain of promises.
If a handler returns a promise (makes another asynchronous request), then the next
handler (success or error) will be called only after that request is finished.
So the previous example code might translate to something like the following, using
promises and the $http service(in AngularJs):
$http.get('/api/server-config').then(
function(configResponse) {
return $http.get('/api/' + configResponse.data.USER_END_POINT);
}
).then(
function(userResponse) {
return $http.get('/api/' + userResponse.data.id + '/items');
}
).then(
function(itemResponse) {
// Display items here
},
function(error) {
// Common error handling
}
);
Propagating Success and Error
Chaining promises is a very powerful technique that allows us to accomplish a lot of
functionality, like having a service make a server call, do some postprocessing of the
data, and then return the processed data to the controller. But when we work with
promise chains, there are a few things we need to keep in mind.
Consider the following hypothetical promise chain with three promises, P1, P2, and P3.
Each promise has a success handler and an error handler, so S1 and E1 for P1, S2 and
E2 for P2, and S3 and E3 for P3:
xhrCall()
.then(S1, E1) //P1
.then(S2, E2) //P2
.then(S3, E3) //P3
In the normal flow of things, where there are no errors, the application would flow
through S1, S2, and finally, S3. But in real life, things are never that smooth. P1 might
encounter an error, or P2 might encounter an error, triggering E1 or E2.
Consider the following cases:
• We receive a successful response from the server in P1, but the data returned is not
correct, or there is no data available on the server (think empty array). In such a
case, for the next promise P2, it should trigger the error handler E2.
• We receive an error for promise P2, triggering E2. But inside the handler, we have
data from the cache, ensuring that the application can load as normal. In that case,
we might want to ensure that after E2, S3 is called.
So each time we write a success or an error handler, we need to make a call—given our
current function, is this promise a success or a failure for the next handler in the promise
chain?
If we want to trigger the success handler for the next promise in the chain, we can just
return a value from the success or the error handler
If, on the other hand, we want to trigger the error handler for the next promise in the
chain, we can do that using a deferred object and calling its reject() method
Now What is deferred object?
Deferred objects in jQuery represents a unit of work that will be
completed later, typically asynchronously. Once the unit of work
completes, the deferred object can be set to resolved or failed.
A deferred object contains a promise object. Via the promise object
you can specify what is to happen when the unit of work completes. You
do so by setting callback functions on the promise object.
Deferred objects in Jquery : https://api.jquery.com/jquery.deferred/
Deferred objects in AngularJs : https://docs.angularjs.org/api/ng/service/$q
then() function is related to "Javascript promises" that are used in some libraries or frameworks like jQuery or AngularJS.
A promise is a pattern for handling asynchronous operations. The promise allows you to call a method called "then" that lets you specify the function(s) to use as the callbacks.
For more information see: http://wildermuth.com/2013/8/3/JavaScript_Promises
And for Angular promises: http://liamkaufman.com/blog/2013/09/09/using-angularjs-promises/
As of ECMAScript6
The .then() method has been included with pure JavaScript with Promises.
From the Mozilla documentation:
The then() method returns a Promise. It takes two arguments: callback
functions for the success and failure cases of the Promise.
The Promise object, in turn, is defined as
The Promise object is used for deferred and asynchronous
computations. A Promise represents an operation that hasn't completed
yet, but is expected in the future.
That is, the Promise acts as a placeholder for a value that is not yet computed, but shall be resolved in the future. And the .then() function is used to associate the functions to be invoked on the Promise when it is resolved - either as a success or a failure.
Before ECMAScript6
To my knowledge, there isn't a built-in then() method in javascript (at the time of this writing).
It appears that whatever it is that doSome("task") is returning has a method called then.
If you log the return result of doSome() to the console, you should be able to see the properties of what was returned.
console.log( myObj.doSome("task") ); // Expand the returned object in the
// console to see its properties.
Here is a thing I made for myself to clear out how things work. I guess others too can find this concrete example useful:
doit().then(function() { log('Now finally done!') });
log('---- But notice where this ends up!');
// For pedagogical reasons I originally wrote the following doit()-function so that
// it was clear that it is a promise. That way wasn't really a normal way to do
// it though, and therefore Slikts edited my answer. I therefore now want to remind
// you here that the return value of the following function is a promise, because
// it is an async function (every async function returns a promise).
async function doit() {
log('Calling someTimeConsumingThing');
await someTimeConsumingThing();
log('Ready with someTimeConsumingThing');
}
function someTimeConsumingThing() {
return new Promise(function(resolve,reject) {
setTimeout(resolve, 2000);
})
}
function log(txt) {
document.getElementById('msg').innerHTML += txt + '<br>'
}
<div id='msg'></div>
Here is a small JS_Fiddle.
then is a method callback stack which is available after a promise is resolved it is part of library like jQuery but now it is available in native JavaScript and below is the detail explanation how it works
You can do a Promise in native JavaScript : just like there are promises in jQuery, Every promise can be stacked and then can be called with Resolve and Reject callbacks, This is how you can chain asynchronous calls.
I forked and Edited from MSDN Docs on Battery charging status..
What this does is try to find out if user laptop or device is charging battery. then is called and you can do your work post success.
navigator
.getBattery()
.then(function(battery) {
var charging = battery.charging;
alert(charging);
})
.then(function(){alert("YeoMan : SINGH is King !!");});
Another es6 Example
function fetchAsync (url, timeout, onData, onError) {
…
}
let fetchPromised = (url, timeout) => {
return new Promise((resolve, reject) => {
fetchAsync(url, timeout, resolve, reject)
})
}
Promise.all([
fetchPromised("http://backend/foo.txt", 500),
fetchPromised("http://backend/bar.txt", 500),
fetchPromised("http://backend/baz.txt", 500)
]).then((data) => {
let [ foo, bar, baz ] = data
console.log(`success: foo=${foo} bar=${bar} baz=${baz}`)
}, (err) => {
console.log(`error: ${err}`)
})
Definition :: then is a method used to solve Asynchronous callbacks
this is introduced in ES6
Please find the proper documentation here Es6 Promises
.then returns a promise in async function.
Good Example would be:
var doSome = new Promise(function(resolve, reject){
resolve('I am doing something');
});
doSome.then(function(value){
console.log(value);
});
To add another logic to it, you can also add the reject('I am the rejected param') call the function and console.log it.
It's about the use of curly braces {} in our arrow functions:
Those 3 examples are doing the same thing (nothing, but have valid grammar, and are a valid Promise chain!)
new Promise(function(ok) {
ok(
/* myFunc1(param1, param2, ..) */
)
}).then(function(){
/* myFunc1 succeed */
/* Launch something else */
/* console.log(whateverparam1) */
/* myFunc2(whateverparam1, otherparam, ..) */
}).then(function(){
/* myFunc2 succeed */
/* Launch something else */
/* myFunc3(whatever38, ..) */
})
console.log("This code has no errors GG!")
The same logic using arrow functions shorthand without {}
new Promise((ok) =>
ok(
/* myFunc1(param1, param2, ..) */
).then(() =>
0 // HEY DID YOU NOTICE! A number that does nothing,
// but otherwise the parsing will fail!
// The code is pretty clean but have a major downside
// As arrow functions without {} can contains only one declaration
// console.log("something") will FAIL here
).then(() =>
"" // HEY DID YOU NOTICE! An empty string that does nothing,
// but otherwise the parsing will fail!
// As arrow functions without {} can contains only one declaration
// We can't add more code here, hence:
// console.log("something")
// Will break the whole promise
// This is likely the error in y(our) code ;)
))
console.log("This code has no errors GG!")
Arrow function with {}
new Promise( (ok) => {
ok(
/* myFunc1(param1, param2, ..) */
)
}).then( () => {
/* myFunc1 succeed */
/* Launch something else */
}).then( () => {
/* myFunc2 succeed */
/* Launch something else */
/* myFunc3(whatever38, ..) */
console.log("something")
/* More console logs! */
console.log("something else")
})
console.log("This code has no errors GG!")
I suspect doSome returns this, which is myObj, which also has a then method. Standard method chaining...
if doSome is not returning this, being the object on which doSome was executed, rest assured it is returning some object with a then method...
as #patrick points out, there is no then() for standard js
doSome("task")must be returning a promise object , and that promise always have a then function .So your code is just like this
promise.then(function(env) {
// logic
});
and you know this is just an ordinary call to member function .
In this case then() is a class method of the object returned by doSome() method.
The ".then()" function is wideley used for promised objects in Asynchoronus programming For Windows 8 Store Apps.
As far as i understood it works some way like a callback.
Find Details in this Documentantion
http://msdn.microsoft.com/en-us/library/windows/apps/hh700330.aspx
Of Cause it could also be the name for any other defined function.
I am about 8 years late, well...anyways, I don't really know what then() does but maybe MDN might have an answer. Actually, I might actually understand it a little more.
This will show you all the information (hopefully), you need. Unless someone already posted this link.
https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Promise/then
The format is promise.prototype.then()
The promise and prototype are kind of like variables but not like variables in javascript, I mean like other things go there like navigator.getBattery().then() where this one actually exists but is barely used on the web, this one shows statuses about the battery of the device, more information and more on MDN if you are curious.