AList.find /AList.tryFind, etc ?

[somelinguist]

Hi, I'm still getting familiar with the library. I was wondering if there is a reason AList.find, AList.tryFind, AList.findIndex, AList.tryFindIndex, etc. weren't part of the library? Maybe also pick/tryPick?

I see that IndexList has implementations for them (except plain pick), but was wondering if there were performance reasons for not having them for AList.

Is there some other way I should use to make an adaptive value from a single item matching a predicate from a clist/alist?

The closest I've come up with so far is AList.filter |> AList.tryFirst, but I'm assuming that goes through the whole list?

Thanks!

[krauthaufen]

Hey,

The things you want are operations on the entire list (at least conceptually) and we simple never had the need for them.
I think they could be implemented "incrementally" s.t. changes won't need to re-evaluate everything.

I'll look into that once i have a little time

[somelinguist]

Thanks for the info. For context, what I'm trying to do have an adaptive value for the currently selected nested item. Something like Sentences[sid].Parts[pid] where sid and pid are themselves adaptive values the represent the index for the currently selected item. sid and pid are really more like keys the way I've implemented them, but they're ordered, which I is why was still looking at using clist/alist instead of cmap/amap. Since they're unique, I'm going through each list, and once I find sid or pid, I know I have the only one.

[krauthaufen]

We use our Index for exactly that, it's comparable and has proper equality and serves as key for alist/clist/etc

[somelinguist]

AreIndexs generic/extendable so that I could I could use my own data type that implements IComparable?

[krauthaufen]

Hey, currently not, but we could theoretically create something like amap using sorted maps instead of hashmaps. I dont thinkt this would be too hard, because all the amap implementations would basically work for that too.
Would that help in your case?

[somelinguist]

Hey. Thanks for the quick reply. That would probably work great. I just need something that can be traversed in a predictable/sorted order and also be used for lookups. I’ve usually just used the find kind of functions on an appropriate ordered collection with some kind of id/key for the lookups, but a sorted/ordered map also sounds very helpful.