Ho to make this call async ```videosList.forEachIn...
# coroutinesk
Kulwinder Singh
12/21/2018, 9:22 AMHo to make this call async
Copy code
videosList.forEachIndexed { index, video: BaseVideo ->
val userData = getUserData(video)*//calling suspend function*
listWithData[index].userData = userData
}g
gildor
12/21/2018, 9:29 AM Copy code
val result: List<UserData> = videosList.map { video ->
async { getUserData(video) }
}.awaitAll()gildor
12/21/2018, 9:29 AMBut in this case you do not limit concurrency
k
Kulwinder Singh
12/21/2018, 9:37 AM"But in this case you do not limit concurrency" h
Kulwinder Singh
12/21/2018, 9:37 AMWhat does this mean , please explain
g
gildor
12/21/2018, 9:37 AMit means that each item of videosList will start own coroutine
gildor
12/21/2018, 9:38 AMit may be fine
gildor
12/21/2018, 9:38 AMdepends on what getUserData does
k
Kulwinder Singh
12/21/2018, 9:38 AMso does it effect perfomance
Kulwinder Singh
12/21/2018, 9:38 AM?
gildor
12/21/2018, 9:38 AMcoroutine itself is cheap.
k
Kulwinder Singh
12/21/2018, 9:39 AMgetUserData just get Document from Firestore collection
g
gildor
12/21/2018, 9:40 AMdon’t know how it works, so have no advices
gildor
12/21/2018, 9:40 AMjust mention that it doesn’t limit concurrency
gildor
12/21/2018, 9:40 AMif you want limit amount of concurrent coroutines you need working pool pattern implementation
k
Kulwinder Singh
12/21/2018, 9:41 AMYes, thanks for link, i will definitely check it out
g
gildor
12/21/2018, 9:44 AMActually, in your case you have also simple way to limit amount of parallel tasks, just use chunked:
Copy code
val result: List<UserData> = videosList.chunked(PARALLEL_COROUTINES_AMOUNT).flatMap { videos ->
videos.map { video ->
async { getUserData(video) }
}.awaitAll()
}k
Kulwinder Singh
12/21/2018, 9:57 AMBut here i know in advance that how many items
videosListchuckedRecyclerViewg
gildor
12/21/2018, 10:02 AMThen don’t worry about concurrency
k
Kulwinder Singh
12/21/2018, 10:11 AMBut still i think that what would happen if user scrolls list lot of time ,then there will be many
asynccoroutinesg
gildor
12/21/2018, 10:37 AMas I said, it depends on getUserData implementation
k
Kulwinder Singh
12/21/2018, 10:51 AM Copy code
private suspend fun getUserData(video: BaseVideo): UserDataAboutVideo? {
return if (currentUser != null) {
val userDataDocument = firestore.getUserDataAboutVideos(currentUser!!.uid)
.document(video.videoId).get().await()
if (userDataDocument.exists())
userDataDocument.toObject(UserDataAboutVideo::class.java)!!
else
null
} else null
}Kulwinder Singh
12/21/2018, 10:51 AMawait implementation is
Copy code
suspend fun <T> Task<T>.await(): T {
return suspendCoroutine { continuation ->
this.addOnCompleteListener { result ->
if (result.result != null && result.isSuccessful)
continuation.resume(result.result!!)
else
continuation.resumeWithException(result.exception?.cause!!)
}
}
}g
gildor
12/21/2018, 10:59 AMI have no idea how Firestore is implemented and work with multiple parallel requests
k
Kulwinder Singh
12/21/2018, 11:12 AMok, No problem, thanks for your all suggestions 👍
Kulwinder Singh
12/21/2018, 11:20 AMyour answers helped me alot
g
gildor
12/21/2018, 2:08 PM👍
11 Views