combine in 0.7

[Alex Cozzi]

Moving my code to 0.7, I used to do, using scalaz:

 val items : DList(Long, (Int, Int)) = ...

 val grouped: DList[(Long, (Int, Int))] = items.groupByKey.combine { _ |+| _ }

But in 0.7 + Scala 2.10 I get an error and I think I need to do:

  val grouped: DList[(Long, (Int, Int))] = items.groupByKey.combine { Sum.int zip Sum.int }

Is there a better/more convenient way to sue Reductions?

[Elliot Chow]

I have been cheating a bit by using an implicit conversion:

implicit def funToReduction[A](f: (A,A) => A) = new Reduction[A]{ val reduce = f }

Would love to know if there is a better way.

[Eric Springer]

I just add a layer of 'Reduction' in, e.g.:

dlist.combine(Reduction(_ |+| _))

But I do feel like having an implicit conversion of (A, A) => A to Reduction[A] would be useful, although I'm not aware of the design goals behind the new combine api, and what it fixes

--
You received this message because you are subscribed to the Google Groups "scoobi-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to scoobi-users...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Eric Springer]

I've now adopted your little trick in my code. It's pretty useful for source-compatibility too, when flipping between 0.6 and 0.7 xD

On Sat, Apr 6, 2013 at 12:39 PM, Elliot Chow <chow....@gmail.com> wrote:

[Eric Torreborre]

Hi Alex,

the Reduction datatype is supposed to provide lots of useful combinators in order to write "combine" functions.

I thought that we had something to create a Reduction from a Semigroup but I couldn't find it so I added it here:

https://github.com/NICTA/scoobi/commit/6505a94f30a0b4c0a34b36be93904a5d0e6a48f1. Your example then becomes:

items.groupByKey.combine(semigroup)

Cheers,

Eric.