Which version would you prefer: a or b? (and why)?

# coroutines

[Paul Woitaschek]

Which version would you prefer: a or b? (and why)?

[altavir]

Why do you need Flow here at all? You do not perform asynchronous operations.

[Paul Woitaschek]

The pagedRecipeIListForIds returns a flow

[Paul Woitaschek]

(you can see the type hints in the first version)

[altavir]

Then you should create a collection, convert it to flow (https://kotlin.github.io/kotlinx.coroutines/kotlinx-coroutines-core/kotlinx.coroutines.flow/kotlin.collections.-iterable/index.html) and then use mapping operations. I am not sure, why you use suspende {} construct

[Paul Woitaschek]

I don't need a Flow<UUID>, i need a List<UUID>

[altavir]

Why?

[Paul Woitaschek]

Because the pageRecipeListForIds requires one

[altavir]

You still convert your list to flow and then use flatmap. You can either use flatmap on a collection and then concat flows, or you can convert a collection to a flow and then work with flows. No principial difference here.

[Paul Woitaschek]

Yes it's a principal difference here

[Paul Woitaschek]

It's the flow flatMap

[Paul Woitaschek]

Not the collection one

[altavir]

I know. It is OK, my correction is only about the generation of initial collection. You can use collection flatmap, get a list of flows and then merge them.
The choice of proper method depends on how does your asynchronous operation works. In general I would prefer to goe with Flow flatMap the way you've done it.

[Paul Woitaschek]

The initial collection is suspending thats the thing here

[Paul Woitaschek]

(should have posted the initial screenshot with the suspending arrow hints in the first place)

[altavir]

Ah, I see. Then the problem is that you have to dive into suspended world. Maybe flowOf(tags). map{recipesByTags...}?

[Paul Woitaschek]

Hah now that is a suggestion 🙂

[altavir]

Because otherwise it is not clear, where do you get a context for collection creation from.