Oh, interesting... and I can see the value in providing an "atom index merge" mechanism.
But, I don't really understand what's going on here: 9 0:`([`2:`]"_)`]} i.3 0 1 0 9 0:`([`1:`]"_)`]} i.3 0 0 2 Thanks, -- Raul On Wed, Jul 3, 2019 at 10:47 AM Henry Rich <henryhr...@gmail.com> wrote: > > 1. x v} y is NOT equivalent to x [`v`]} y. The difference is that the > former treats the result of the verb as atom numbers, the latter as item > numbers. Different parts of the array are being modified if items are > not atoms, i. e. if rank>1. > > 2. Already in x gerund} y you have the possibility that v1 can return > another gerund in the execution of (x v0 y) (x v1 y)} x v2 y. For the > record, I wanted to disallow this frivolity with a message '0C3 abend', > but was shouted down by users. > > 3. Enhancements to x v} y should not be expected. > > Henry Rich > > On 7/3/2019 10:34 AM, Raul Miller wrote: > > Note also that if V is a verb which selects indices, X is new values > > and Y is an array to merge, then: > > > > X V} Y > > is equivalent to > > X [`V`]} Y > > or to > > X [`(V)`]} Y > > > > It's not as concise, though. And, hypothetically, we might like to have: > > > > merge=:2 :0 > > : > > G=. X V Y > > X G} Y > > ) > > > > instead (in other words, it might be nice if X V} Y would have V > > compute a gerund to be consumed by the } conjunction). > > > > That said, since a gerund is never indices, I imagine that if we had a > > use for this form, we could achieve that without having to sacrifice > > the current behavior... > > > > Thanks, > > > > > --- > This email has been checked for viruses by AVG. > https://www.avg.com > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
