Hi, I am new to D and would like to know if there is any built in feature for accessing a subarray (uni or multi dimensional) of an array. For example let

arr = [[1,2,3,4,5,6],[7,8,9,10,11,12],[13,14,15,16,17,18]]

I want to obtain a subarray like

assert(arr[][0] == [1,7,13]);
assert(arr[1][] == [7,8,9,10,11,12];
assert(arr[0,2][1] == [2,14]);
assert(arr[0,2][0..3,5] == [[1,2,3,6],[13,14,15,18]]);

Thank you.

Omid:

> arr = [[1,2,3,4,5,6],[7,8,9,10,11,12],[13,14,15,16,17,18]]
>
> I want to obtain a subarray like
>
> assert(arr[][0] == [1,7,13]);

This gives a range, not an array, so if you want an array, you have to call std.array.array later:

http://dlang.org/phobos/std_range.html#transversal


> assert(arr[1][] == [7,8,9,10,11,12];

Just use array[1]. Or array[1].dup is you want a copy.


> assert(arr[0,2][1] == [2,14]);

Enumerated indexes are not supported, so you need a iota + filter + transversal + array.


> assert(arr[0,2][0..3,5] == [[1,2,3,6],[13,14,15,18]]);

This requires a iota + filter + chain + array, or something like that.

If you don't want to do all this, you have to write your slicing and dicing array struct :-( Someday it will probably be present in Phobos.

Bye,
bearophile

On Sunday, 30 December 2012 at 12:51:58 UTC, bearophile wrote:
> Omid:
>
>> arr = [[1,2,3,4,5,6],[7,8,9,10,11,12],[13,14,15,16,17,18]]
>>
>> I want to obtain a subarray like
>>
>> assert(arr[][0] == [1,7,13]);
>
> This gives a range, not an array, so if you want an array, you have to call std.array.array later:
>
> http://dlang.org/phobos/std_range.html#transversal
>
>
>> assert(arr[1][] == [7,8,9,10,11,12];
>
> Just use array[1]. Or array[1].dup is you want a copy.
>
>
>> assert(arr[0,2][1] == [2,14]);
>
> Enumerated indexes are not supported, so you need a iota + filter + transversal + array.
>
>
>> assert(arr[0,2][0..3,5] == [[1,2,3,6],[13,14,15,18]]);
>
> This requires a iota + filter + chain + array, or something like that.
>
> If you don't want to do all this, you have to write your slicing and dicing array struct :-( Someday it will probably be present in Phobos.
>
> Bye,
> bearophile

Thanks, though I didn't understand a word you said:). I should learn iota, filter, transversal first.

Omid:

> Thanks, though I didn't understand a word you said:).

This code works with the latest beta compiler:


import std.algorithm: filter, equal, map, canFind;
import std.range: transversal, iota;

void main() {
    auto arr = [[ 1, 2, 3, 4, 5, 6],
                [ 7, 8, 9,10,11,12],
                [13,14,15,16,17,18]];

    auto r1 = transversal(arr, 0);
    assert(r1.equal([1, 7, 13]));

    assert(arr[1] == [7,8,9,10,11,12]);

    auto r2 = [0, 2]
              .map!(i => arr[i])()
              .transversal(1);
    assert(r2.equal([2, 14]));

    auto r3 = [0, 2]
              .map!(i => arr[i][0..3] ~ arr[i][5])();
    assert(r3.equal([[1,2,3,6],[13,14,15,18]]));
}


Bye,
bearophile

On Sunday, 30 December 2012 at 18:55:50 UTC, bearophile wrote:
> Omid:
>
>
> This code works with the latest beta compiler:
>
>
> import std.algorithm: filter, equal, map, canFind;
> import std.range: transversal, iota;
>
> void main() {
>     auto arr = [[ 1, 2, 3, 4, 5, 6],
>                 [ 7, 8, 9,10,11,12],
>                 [13,14,15,16,17,18]];
>
>     auto r1 = transversal(arr, 0);
>     assert(r1.equal([1, 7, 13]));
>
>     assert(arr[1] == [7,8,9,10,11,12]);
>
>     auto r2 = [0, 2]
>               .map!(i => arr[i])()
>               .transversal(1);
>     assert(r2.equal([2, 14]));
>
>     auto r3 = [0, 2]
>               .map!(i => arr[i][0..3] ~ arr[i][5])();
>     assert(r3.equal([[1,2,3,6],[13,14,15,18]]));
> }
>
>
> Bye,
> bearophile

Thanks again. I started learning std.array and I should say it fits all my needs as a mathematics and combinatorics researcher. Now I think I am able to give up commercial softwares.

On Monday, 31 December 2012 at 13:29:02 UTC, Omid wrote:

>
>I started learning std.array

Actually, std.array and std.range both.