id like to also ask how to rewrite this in more fu...

# getting-started

[ursus]

id like to also ask how to rewrite this in more functional style

val maxSize = BookDao.RECENTS_SIZE
val all = mutableListOf<RealmObject>()
all.addAll(it.findAllRecentSearchedBooks().run { if (this.size > maxSize) subList(0, maxSize) else this })
all.addAll(it.findAllRecentSearchedAuthors().run { if (this.size > maxSize) subList(0, maxSize) else this })

all.sortByDescending {
	when (it) {
	        is RecentlySearchedBook -> it.timestamp
	        is RecentlySearchedAuthor -> it.timestamp
	        else -> error("Unknown type")
        }
}

val allUnmamaged = mutableListOf<SearchItem>()
for (i in 0 until Math.min(all.size, maxSize)) {
	allUnmamaged.add(it.unmanaged(all[i]) as SearchItem)
}

allUnmamaged

[diesieben07]

val all = it.findAllRecentSearchedBooks().take(maxSize) + it.findAllRecentSearchedAuthors.take(maxSize)
          .sortedByDecending {...}

For a start.

[diesieben07]

and then you can do

val allUnmanaged = all.map { item -> it.unmanaged(item) as SearchItem }

[ursus]

sweet

[ursus]

is there a way to not break the chain, i.e. sort returns unit

[diesieben07]

sorted returns sorted copy, sort sorts in place

[ursus]

i see, thanks

[diesieben07]

So if you want to avoid all the intermediate collections, use sequences

[ursus]

arent sequences from coorutines?

[diesieben07]

But if that is needed depends on the size of your collections

[diesieben07]

No

[diesieben07]

Standard library

[ursus]

ill look into it, thnx

[diesieben07]

https://kotlinlang.org/api/latest/jvm/stdlib/kotlin.sequences/index.html

[diesieben07]

like an iterable, but lazy. So if you map twice with a sequence, your collection is only traversed once

[ursus]

you mean instantiated?

[diesieben07]

No. If you have a List and you call map on it, a new List of the same length will be allocated and the result will be stored in it.

[diesieben07]

So if you map twice, you will get two lists

[diesieben07]

With a Sequence you still get new objects allocated, but they only store the source collection and the operation

[ursus]

yea, thats what I meant

[diesieben07]

And the operations are all applied in one go if you convert to a normal collection again in the end

[diesieben07]

Similar to a Stream in the JDK

[ursus]

so its like inline imperative collection processing with functional api?

[diesieben07]

Sorry, I have no idea what you mean by that 😄

[ursus]

like, gc aware functional style

[ursus]

or whatever :d

[diesieben07]

I guess you could call it that. Like I said, depends on how big your collections are

[diesieben07]

For small collections the non-lazy approach is faster

[diesieben07]

You gotta benchmark for your specific use case if you have a problem

[ursus]

k, thanks