How do I get an array from randomSample()?

int[] source = [ ... ];
int[] sample = randomSample(source, 3);

src\main.d(30): Error: cannot implicitly convert expression (randomSample(source, 3u)) of type RandomSample!(int[], void) to int[]

I get that RandomSample is a struct which implements the necessary interface to work with foreach.  But the fact that this struct is considered a proprietary implementation detail of std.random is a constant source of frustration for me.  Quite a few D libraries use this trick, but the opacity of these data structures make it impossible for users to know exactly how to use them outside of the one or two examples given in the documentation.  I don't know how to improve the situation, other than documenting them explicitly rather than treating them as a magical black box.

In the mean time, can anyone suggest a solution to the above, short of actually iterating over it?  I suppose someone will say to use map!().

Dave

On 5/17/2014 9:18 PM, David Held wrote:
> How do I get an array from randomSample()?
>
> int[] source = [ ... ];
> int[] sample = randomSample(source, 3);
>
> src\main.d(30): Error: cannot implicitly convert expression
> (randomSample(source, 3u)) of type RandomSample!(int[], void) to int[]
> [...]

Even worse, foreach is brittle:

    foreach (i, a; randomSample(source, 3))

src\main.d(30): Error: cannot infer argument types

If I remove the index variable, it works.  But I really need the index too.

Dave

On Sunday, 18 May 2014 at 04:19:05 UTC, David Held wrote:
> How do I get an array from randomSample()?
>
> int[] source = [ ... ];
> int[] sample = randomSample(source, 3);
>
> src\main.d(30): Error: cannot implicitly convert expression (randomSample(source, 3u)) of type RandomSample!(int[], void) to int[]
>
> I get that RandomSample is a struct which implements the necessary interface to work with foreach.  But the fact that this struct is considered a proprietary implementation detail of std.random is a constant source of frustration for me.  Quite a few D libraries use this trick, but the opacity of these data structures make it impossible for users to know exactly how to use them outside of the one or two examples given in the documentation.  I don't know how to improve the situation, other than documenting them explicitly rather than treating them as a magical black box.
>
> In the mean time, can anyone suggest a solution to the above, short of actually iterating over it?  I suppose someone will say to use map!().
>
> Dave

You need to use the function array from std.array.

import std.array;

int[] source = [ ... ];
int[] sample = randomSample(source, 3).array();

Array has the ability to turn any range into an array, and it's generally how you convert from these structs (which generally implement a range interface) returned by most range-based functions in D. It can be a bit confusing at first, but returning a range struct is much more flexible than returning a simple array of values.

Meta:

> You need to use the function array from std.array.
>
> import std.array;
>
> int[] source = [ ... ];
> int[] sample = randomSample(source, 3).array();

In some cases it's also useful to use std.algorithm.copy:


void main() {
    import std.stdio, std.algorithm, std.random, std.array,
           std.range;

    immutable int[9] source = iota(10, 100, 10).array;
    source.writeln;
    int[3] sample;
    source[].randomSample(sample.length).copy(sample[]);
    sample.writeln;
}


Bye,
bearophile