slick... virtually identical times and yours is more general and clear.
thanks
(6!:2) 'idotfill (>;:''SF SY PK ... '');(5e6#(>''SF'';''SY''))'
0.150771
(6!:2) '(>;:''SF SY PK ... '') chidot (5e6#(>''SF'';''SY''))'
0.158705
On Wed, Apr 11, 2018 at 4:44 PM, Raul Miller <rauldmil...@gmail.com> wrote:
> Here's what I'd start with:
>
> chidot=: #@[ ({. i. }.) ,
>
> If that was too slow, I'd probably build an explicit work-alike, with
> intermediate variables for shapes.
>
> (>;:'SF SY PK ...') chidot (>;:'SF SY')
> 0 1
>
>
> --
> Raul
>
> On Wed, Apr 11, 2018 at 4:35 PM, Joe Bogner <joebog...@gmail.com> wrote:
> > I often get tripped up with different trailing shapes when using i. with
> > char arrays. I must not be the only one[1]
> >
> > Does this look like a suitable utility to deal with it or is there a more
> > common way?
> >
> > Y1=:0&{::
> > Y2=:1&{::
> > idotfill =. (( ( (_,]) {. Y1@[) i. ((_,]) {. Y2@[) ) {:@$@Y2 >.
> {:@$@Y1) f.
> >
> > idotfill (>'SF';'SY';'PK ');(>'SF';'SY')
> > 0 1
> >
> > compare with
> >
> > NB. different trailing shape
> > (>'SF';'SY';'PK ') i. (>'SF';'SY')
> > 3 3
> >
> > NB. remove the trailing space on PK
> >
> > (>'SF';'SY';'PK') i. (>'SF';'SY')
> > 0 1
> >
> >
> > [1] - "y must match the entire row of x (including any trailing fill
> > letters) for (i.) to find it. Otherwise (i.) returns (#x) (signifying
> "not
> > found"), concealing your error."
> >
> > http://code.jsoftware.com/wiki/Vocabulary/idot#dyadic
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
