Actually, my approach did not properly handle cases which should give
an empty result or length 1 result.
Fixed version:
an0=: 2#0~:_2 +/\ ]
fix=: ^:(1<#)
an1=: ({.,(#~ an0)@}.@}:,{:)fix@(#~ an0)fix
an2=: an1^:_
#an2 1 _1 1 _1
0
an2 3
3
Thanks,
--
Raul
On Fri, Dec 6, 2019 at 2:24 PM Raul Miller <[email protected]> wrote:
>
> Here's one approach:
>
> an0=: 2#0~:_2 +/\ ]
> an1=: ({.,(#~ an0)@}.@}:,{:)@(#~ an0)
> an2=: an1^:_
> an2 1 _1 2 _2{~100?.@#4
> 2 1 2 1 _2 _1 2 1 1 2 2 2 _1 2 _1 2 1 2 2 2 1 1 _2 _1 2 1 _2 _1 _2 _2
> _2 _2 _1 _2 _1 _1 _2 _2 1 1 1 2 1 _2 1 _2 _1 _1 _1 2
> #an2 1 _1 2 _2{~100?.@#4
> 50
>
> I have not convinced myself that there's no ambiguity in this problem,
> and my approach seems like it has room for improvement, but I get the
> same result you got.
>
> Thanks,
>
> --
> Raul
>
> On Fri, Dec 6, 2019 at 7:15 AM R.E. Boss <[email protected]> wrote:
> >
> > Given an array with zero or more annihilating pairs, i.e., two subsequent
> > numbers which add up to zero, the question is to clean up the array by
> > deleting all annihilating pairs such that no such pairs are left.
> > I do have a solution that is both elegant and efficient (I believe), but I
> > am curious about other thoughts.
> >
> > foo 2 1 1 _1 2 _2 _1
> > 2
> >
> > foo 1 _1 2 _2{~100?.@#4 NB. Notice (?.)
> > 2 1 2 1 _2 _1 2 1 1 2 2 2 _1 2 _1 2 1 2 2 2 1 1 _2 _1 2 1 _2 _1 _2 _2 _2 _2
> > _1 _2 _1 _1 _2 _2 1 1 1 2 1 _2 1 _2 _1 _1 _1 2
> >
> >
> > R.E. Boss
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
