Type mismatch error when call map() method on a map: case vs function literal

Hi guys,

I'm reading Programming in Scala and trying some code examples in the
book. There is a code snippet like:
  val m = Map("a" -> 1)
  m.map { case (k, v) =>  (k + "!", "x" * v)}
This works. Then I tried to fool around a little bit by changing the
case in the curly braces to a function literal.
  m.map((k: String, v: Int) => (k + "!", "x" * v))
Now I get a compiling error:
  type mismatch;  found   : (String, Int) => (String, String)
required: (String, Int) => ?

I don't understand what's the error means. Can anybody tell me why I
can't use function literal here?

I'm using Scala IDE 2.0 on Eclipse 3.7. Scala version is 2.9.1

Re: Type mismatch error when call map() method on a map: case vs function literal

I figured it out, the correct way to use function literal here is:
  m.map((t: (String, Int)) => (t._1 + "!", "x" * t._2))
Because map() method takes one Tuple2[String, Int] as parameter, not
two parameters(a String and an Int)
I think the error message is a little misleading, though.
On Jan 1, 11:38 am, BlackFish <qsha...@...> wrote:
> Hi guys,
>
> I'm reading Programming in Scala and trying some code examples in the
> book. There is a code snippet like:
>   val m = Map("a" -> 1)
>   m.map { case (k, v) =>  (k + "!", "x" * v)}
> This works. Then I tried to fool around a little bit by changing the
> case in the curly braces to a function literal.
>   m.map((k: String, v: Int) => (k + "!", "x" * v))
> Now I get a compiling error:
>   type mismatch;  found   : (String, Int) => (String, String)
> required: (String, Int) => ?
>
> I don't understand what's the error means. Can anybody tell me why I
> can't use function literal here?
>
> I'm using Scala IDE 2.0 on Eclipse 3.7. Scala version is 2.9.1

Re: Re: Type mismatch error when call map() method on a map: case vs function literal

Re: Type mismatch error when call map() method on a map: case vs function literal

Thanks for your thorough explanation,  Doug.

On Jan 1, 12:23 pm, Doug Tangren <d.tang...@...> wrote:
> On Sat, Dec 31, 2011 at 11:12 PM, BlackFish <qsha...@...> wrote:
> > I figured it out, the correct way to use function literal here is:
> >  m.map((t: (String, Int)) => (t._1 + "!", "x" * t._2))
> > Because map() method takes one Tuple2[String, Int] as parameter, not
> > two parameters(a String and an Int)
> > I think the error message is a little misleading, though.
>
> When you call map on a Map, a seq of tuples, you are applying a function to
> equals tuple
>
> { case (k, v) =>  (k + "!", "x" * v) } is a partial function that is always
> defined for a 2 item tuple
>
> The `case` keyword here applies the a tuple extractor on the map
> entires, (k, v)
>
> to get the same effect using a function literal you can use
>
> m.map((t: (String, Int)) => t match { case (k,v) => (k + "!", "x" * v) })
>
> or drop the type annotations, for less verbosity
>
>  m.map(t => t match { case (k, v) => (k + "!", "x" * v) })
>
> or use the _ shorthand, for even less verbosity
>
> m.map(_ match { case (k,v) => (k + "!", "x" * v) })
>
> or the simplest way is just to use the partial function directly, as
> mentioned above
>
> m.map({ case (k,v) => (k + "!", "x" * v) })

Re: Re: How to avoid Int and Double boxing in a returned tuple ?

Re: Immutable graph libraries and alternative structures

Hi Tim,

>I know there has been more than one go at writing graph libraries.
Do you mean Scala graph libraries? As I began work on Graph4Scala
roughly one year ago, EPFL backed the idea of my contribution because
there was no general purpose Scala alternative.

>… I need immutable graph structures of fairly modest size (a few hundred to about 4000 nodes)… they are
typically highly  disconnected… The graphs are also directed and not cyclic. So I'm looking for a
library that will support that sort of structure
efficiently and immutably - though I'm not unhappy doing one by hand.

Graph4Scala does meet your above requirements. If you wanted even
automatic maintenance of acyclicity for your graph instances you could
rely on this feature of the next Graph4Scala release. But I assume
that’s not your point.

Certainly you may have a lot of other, including non-technical reasons
to prefer an individual implementation(?).

If you still considered Graph4Scala, the next step would be to make
thoughts about how to integrate your algorithms and which node/edge
layout to select to achieve an optimal solution with minimal efforts.

Peter

Reflection

Hi
I want to use reflection to generate case classes from XML.
Any recommendations for a reflection tool for the job?
Can't use 2.10 nightly builds and new API because of coordination with
other libraries.

Tim

Re: Reflection

What's wrong with Class.forName? It has the virtue of simplicity, even
if it isn't necessarily the prettiest call in the standard library.

Donald

On Sun, Jan 1, 2012 at 9:10 AM, Tim P <tim.pigden@...> wrote:
> Hi
> I want to use reflection to generate case classes from XML.
> Any recommendations for a reflection tool for the job?
> Can't use 2.10 nightly builds and new API because of coordination with
> other libraries.

REPL question

Hi,

I am often writing small scala programs using Emacs and the REPL.
However, whenever
I define something like

abstract class Rexp {
  def ~ (right : Rexp) = SEQ(this, right)
  def || (right : Rexp) = ALT(this, right)
  def ** = STAR(this)
}

case object NULL extends Rexp
case object EMPTY extends Rexp
case class CHAR(c: Char) extends Rexp
case class ALT(r1: Rexp, r2: Rexp) extends Rexp
case class SEQ(r1: Rexp, r2: Rexp) extends Rexp
case class STAR(r: Rexp) extends Rexp

then the REPL complains that the value for SEQ and ALT is not found in
the
abstract class. The compiler (scala) accepts such definitions. Is
there an elegant
way to make the REPL also accept such definitions? Of course I can
wrap
everything into an object and everything in a big chunk, but I hope
there is
a better solution.

Thanks for the help,
Christian

Re: REPL question