This is relevant and interesting: http://stackoverflow.com/questions/16108899/understanding-j-array
On Tue, Apr 23, 2013 at 5:49 PM, Raul Miller <[email protected]> wrote: > When I read http://www.jsoftware.com/help/dictionary/d210.htm I see the > text: > > "The fit conjunction ($!.f) provides fill specified by the atom f , or > the normal fill defined under Take ({.) if f is an empty vector." > > So I am not sure why this behavior would be unexpected. > > -- > Raul > > On Mon, Apr 22, 2013 at 9:42 PM, Henry Rich <[email protected]> wrote: > > I'm not talking about > > > > 3 $ 4 > > 4 4 4 > > > > or even > > 3 $ i. 2 3 > > 0 1 2 > > 3 4 5 > > 0 1 2 > > > > What happens when the right argument has no atoms? Like > > > > 3 $ (0$0) > > ? > > > > > > > > 3 $ (0$0) > > |length error > > > > You can see the problem. Since all empties are considered identical, > what > > value should Roger make 3 copies of? > > > > But... do you have to make the right operand non-empty, or is there a > way > > to use $ with empty y? > > > > > > > > There is... > > > > 3 $!.5 (0$0) > > 5 5 5 > > > > You can even keep the type of the empty: > > > > 3 $!.'' (0$0) > > 0 0 0 > > 3 $!.'' (0$a:) > > ++++ > > |||| > > ++++ > > > > As I read the Dictionary, this behavior is unexpected - but welcome. > > > > Henry Rich > > > > ---------------------------------------------------------------------- > > 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
