https://issues.dlang.org/show_bug.cgi?id=6788

--- Comment #12 from greenify <greeenify@gmail.com> ---
@andralex: I think there is also a decision-blocked PR pending in the Phobos queue:

https://github.com/dlang/phobos/pull/4027

--

https://issues.dlang.org/show_bug.cgi?id=6788

--- Comment #13 from Andrei Alexandrescu <andrei@erdani.com> ---
I see in https://docs.python.org/2/library/itertools.html that the recipe given
for pairwise would produce the pairs (1, 2), (2, 3), and (3, 4) starting from
[1, 2, 3, 4]. So defining it with different semantics is potentially confusing.

The functionality proposed is really "combinations of 2 elements in the given range". In general we'd be looking at combinations of k elements of the given range, yielding a range of k-tuples. See https://en.wikipedia.org/wiki/Combination

Fittingly enough, Python offers https://docs.python.org/2/library/itertools.html#itertools.combinations and https://docs.python.org/2/library/itertools.html#itertools.combinations_with_replacement

I think ws should define a function as follows:

auto combinations(uint k, Flag!"Duplicates" f = Duplicates.no, R)(R range);

The function returns a range of k-tuples containing combinations of elements in the range. Depending on the flag, duplicates are present or not.

--

https://issues.dlang.org/show_bug.cgi?id=6788

Seb <greensunny12@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |greensunny12@gmail.com
            Summary|std.range.pairwise          |std.algorithm.combinations

--- Comment #14 from Seb <greensunny12@gmail.com> ---
> I see in https://docs.python.org/2/library/itertools.html that the recipe given for pairwise would produce the pairs (1, 2), (2, 3), and (3, 4) starting from [1, 2, 3, 4]. So defining it with different semantics is potentially confusing.


Yes, especially now that we have pairwise as slides.

> I think ws should define a function as follows:
> auto combinations(uint k, Flag!"Duplicates" f = Duplicates.no, R)(R range);

> The function returns a range of k-tuples containing combinations of elements in the range. Depending on the flag, duplicates are present or not.

+1 (renamed accordingly)


@ students - see also: http://docs.mir.dlang.io/latest/mir_combinatorics.html (this is my module, it can be ported)

--

https://issues.dlang.org/show_bug.cgi?id=6788

--- Comment #15 from Seb <greensunny12@gmail.com> ---
*** Issue 13039 has been marked as a duplicate of this issue. ***

--

https://issues.dlang.org/show_bug.cgi?id=6788

tyckesak <josipp@live.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
                 CC|                            |josipp@live.de
         Resolution|---                         |DUPLICATE

--- Comment #16 from tyckesak <josipp@live.de> ---
Obviated by the introduction of `std.algorithm.setops.cartesianProduct` into the standard library.

*** This issue has been marked as a duplicate of issue 7128 ***

--