On 8/23/07, Raul Miller <[EMAIL PROTECTED]> wrote: > > ins=:1 :0 > > : > > b=.1#~1 j.m e.~i.#y > > x (I.-.b)} b expand y > > )
Brian Schott pointed out to me that this (and my variant ins1) assume that m is in ascending order (which I was assuming, since I don't normally do this kind of thing with indices but with bit masks). Here's a version which corrects this problem: ins=:1 :0 : b=.1#~1 j.m e.~i.#y (x/:m) (I.-.b)} b expand y ) Thanks, -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
