Arie,
I now see that my comment regarding a Hook was irrelevant (if not just
wrong) because the lha of the amend gerund supplies the m in m}. So to get
different strides is just a matter of supplying a different set of array
indices for the lha From. So, all that's needed for different strides is
Yes it's Arie = Aai = @@i (sounds like I). :-)
On 20-02-19 16:57, Brian Schott wrote:
Aii,
(Is it Ari?)
Yes, that is a nice extension that absolutely anticipates my expanded
needs. I played with increasing the dimensions of i and w a little and it
works beautifully.
(Please follow below this
Aii,
(Is it Ari?)
Yes, that is a nice extension that absolutely anticipates my expanded
needs. I played with increasing the dimensions of i and w a little and it
works beautifully.
(Please follow below this little demonstration of expanded dimensions for
more discussion)
;/($ $"1 ((0 0;0 _1;_
Perhaps this is also useful:
ka=:'abcdefghijklmnopqrstuvwxyz'
Ga=:'MNOPQRSTUVWXYZABCDEFGHIJKL'
NB. stretch -- amend -- unstretch (reshape)
($ $"1 ((0 0;0 _1;_1 0 ;_1 _1){($ w),;._3 i. $ i)(,w)"_`[`] }"1 ,) i=.3
3 $ Ga [ w=. 2 2 $ ka
Handles variable 2D shaped arrays AFAICS
Also of course:
You are right, this could have been done using the gerund form of }
w=: 2 2$'abcd'
i=: 3 3$'MNOPQR'
w"_`]`(i"_)}~@<"1(#:i.4){;/+/~0 1
Basically,
w"_`]`(i"_)}~
behaves like
(w 2 :'m y}n'i)
They are not equivalent in all contexts (they have different display
forms, for example), bu
Raul,Yes. That is TOO clever (see why, below).Your approach provides a
way to avoid using gerund form of } .Making the <"1 part of a verb
(and especially the "1 part of <"1) instead of including it in the
construction of the verb's righthand argument noun, brings your
conjunction's y into the amend
Mine wasn't a solution, more of a hint. I got the rank wrong.
That said, the issue you mentioned would break code in just about any
language: my quotes had gotten changed to something else.
Anyways, here's a solution:
(w 2 :'m y}n'i)@<"1(#:i.4){;/+/~0 1
abO
cdR
MNO
Mab
Pcd
MNO
MNO
abR
cdO
Oops. I should have pointed out that Ric's result does have one less layer
of boxing.
On Tue, Feb 19, 2019 at 12:31 AM Brian Schott
wrote:
> Bob and Ric,
>
> Focusing separately on the 3 parts of the gerund form by placing ; between
> the parts, really helps me see what you have done. (And btw,
Bob and Ric,
Focusing separately on the 3 parts of the gerund form by placing ; between
the parts, really helps me see what you have done. (And btw, i should have
contained 'STU' also. Thanks, again.
Bob's version and Ric's version both produce virtually identical results; I
prefixed Bob's version
Nice Ric,
I like the way that you packaged the selector inside the amend clause, as that
gets around a lot of the boxing that I needed to coordinate the arguments.
The only improvement I can see is that the infinite rank on amend is not needed
as the rank 2 covers the w ; i argument. At least t
Here is a version of the same idea but with less boxing. Interestingly it
looks as though J9 will support the specifying indices using unboxed arrays
which would simplify this further!
load 'stats'
((0 1,:1 2) {~ permrep 2) (0&{::@]) `(<@:(<"1)@[)`(1&{::@]) }"2 _ w;i
abO
cdR
MNO
Mab
Pcd
M
Hi Brian,
My comments about boxing had more to do with all the boxing that needed to be
done and then undone to get the right result.
A couple of examples.
(<"0 ind) to me just looks messy especially when it is already boxed, but the
each/every adverb requires that.
(< cmb) to just unpack i
Bob,
Yes. That's what I wanted.
I'm not completely sure about your comment about the necessity of boxing
and unboxing, though.
I changed your each to every and got an unboxed result just like the result
you gave, but I think your comment is deeper and I want to think more about
it.
I especially w
I'll study Raul's solution a bit more, but in the meantime I think that this is
what you may be looking for:
[combis =: #:i. 4 NB. same definitions as before
0 0
0 1
1 0
1 1
[w=: 2 2$'abcd'
ab
cd
[i=: 3 3$'MNOPQR'
MNO
PQR
MNO
[ind=: combis<"1@:{2<\i. 3 NB. capture the indices that
You have a bunch of different values for r, and the final value has
everything already “in place”.
So it’s not completely clear what problem you are trying to solve.
That said, my best guess would be that you are trying to achieve “use rank
conjunction on the adverb argument to amend.”
If so, th
The snippet below produces a value for r which I would like to produce more
elegantly: without the for. loop suggested by the repeated building of
r-values. Perhaps a gerund-powered amend could work, but I am not
optimistic enough to travel there without some encouragement.
The result in r is atte
16 matches
Mail list logo
