Raul -
NuVoc's "Atop" page states: "... The result of u@v is a tacit verb
equivalent to [x] (u@:v)"v y ..."
So, using your last example, and as (#:) has dyadic rank (1 0),
(1 {"1) and (#:) join flawlessly
3 5 5 (1 & {"1 @ #:) 13 17 NB. verbose
2 3
3 5 5 (1 & { @ #:) 13 17 NB. compact
2 3
and as the compound (u @ v) inherits the rank of v, it still accepts
a vector on its left while working on single elements on the right:
3 5 5 (1 & { @ #:"1 0) 13 17 NB. verbose
2 3
3 5 5 (1 & { @ #:)"1 0 (13 17) NB. verbose
2 3
3 5 5 (1 & { @ #:) 13 17 NB. compact
2 3
I think that covers it.
Thanks
-M
At 2016-06-14 21:49, you wrote:
There is something wrong with how your email client quotes my email,
or how my email system sends them to you, so I am going to trim
quotes and reply inline here. On Tue, Jun 14, 2016 at 12:34 PM,
Martin Kreuzer <[email protected]> wrote: > Your next to last
sentence reads, I quote, > "Put differently, we do not need the "1
(from 1 {"1 y) because we get an > implicit "0 from our use of @ (in
1&{@#:)" > > Q: Should this read "... because we get an ** implicit
"1 ** from our use of > ..."..? > If not, then I'm most probably
still missing something. Consider this example: 3 5 5 #: 13 17 0
2 3 0 3 2 Here, the number 13 corresponds to the row 0 2 3, and the
number 17 corresponds to the row 0 3 2. So if we deal with the
result separately, we need to pull the second column from rank 1
(rows), while if we build a verb that does both steps together, that
verb will be dealing with rank 0 (individual numbers). Does that
help put things in perspective? Thanks, -- Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
