FWIW: Catching and replacing 'threatening' values:
(at the cost of performance and elegance)
{&'#.' @ (2:<|) @ ((_:`(+*:)@.(_~:|@]))"0^:400 0:) (18 %~ i:_20) j.~/
28 %~ _59+i.75
a bit more speed:
{&'#.' @ (2:<|) @ (([(+*:)_[`(I.@:(_=|@]))`]}"1 ])^:400 0:) (18 %~
i:_20) j.~/ 28 %~ _59+i.75
Hallo June Kim, je schreef op 31-05-09 13:09:
> Running the following code on J602 from
> http://web.archive.org/web/20070407115228/www.ewartshaw.co.uk/jwhat.html
>
> {&'#.' @ (2:<|) @ ((+*:)^:400 0:) (18 %~ i:_20) j.~/ 28 %~ _59+i.75
> |NaN error
> | {&'#.'@(2:<|)@((+*:)^:400 0:)(18%~i:_20)j.~/28%~_59+i.75
>
> It runs okay on J601. I guess the core of the difference is:
>
> *:__j__
> _
>
>
> vs.
>
>
> *:__j__
> |NaN error
> | *:__j__
>
> (the former is from J601, and the other from J602)
>
> How could I change the code to handle it and produce the originally
> intended graph?
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
=@@i
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
