Perhaps you're misunderstanding how squad indexing works?
⎕←t←5 5⍴⎕A
ABCDE
FGHIJ
KLMNO
PQRST
UVWXY
(1 3 5)(1 3 5)⌷t
ACE
KMO
UWY
t[1 3 5;1 3 5]≡(1 3 5)(1 3 5)⌷t
1
Perhaps you wanted to use something like Dyalog's choose indexing?
x[(∊2=⍴¨x)/,s]
┌──┬──┐
│XX│XX│
└──┴──┘
Jay.
On Mon, 5 Nov 2018 at 13:07, Juergen Sauermann <
[email protected]> wrote:
> Hi Hans-Peter,
>
> I tried to check your example with IBM APL2 and it
> gives me a RANK ERROR on
>
> ((∊2=⍴¨x)/,s)⌷x
>
> ⍴((∊2=⍴¨x)/,s) is 3 and ⍴x is 4 4.
> I have earlier noticed some differences in nesting levels between different
> APL interpreters that
> could be traced back to what the ISO standard calls "reduction styles" (an
> implementation parameter).
> I suppose the reduction style also matters for the EACH operator
>
> For example, GNU APL and IBM APL2 use the "Enclose-Reduction-Style" while
> Dyalog APL uses
> the "Insert-Reduction-Style". So I wonder where your examples originate.
>
> It would help me a lot if you could break down your examples into individual
> steps like
> *4⎕CR y1←⍴¯x**
> 4⎕CR y2←2=y1
> **4⎕CR y3←∈y2**4⎕CR y**3**/,s**...*
> so that I can see the expected intermediate results (and a hint which ones
> are wrong)
>
> Best Regards,
> /// Jürgen Sauermann
>
>
> On 11/05/2018 12:49 AM, Hans-Peter Sorge wrote:
>
> Hi Jürgen,
>
> it looks like there is some inconsistency.
>
> x←4 4⍴'XX' '' '' '' '' '' '' ''
>
> s←(⍳4)∘.,⍳4
> s
> 1 1 1 2 1 3 1 4
> 2 1 2 2 2 3 2 4
> 3 1 3 2 3 3 3 4
> 4 1 4 2 4 3 4 4
>
>
> ((∊2=⍴¨x)/,s)⌷x
> XX
> XX
> ----- expected (no leading empty elements)
> XX
> XX
>
> ⍴((∊2=⍴¨x)/,s)⌷x
> 2 2 ----- expected 2 1 (only one column)
> ⍴¨((∊2=⍴¨x)/,s)⌷x
> 0 2
> 0 2
> ----- expected (no leading empty elements)
> 2
> 2
> -------------------------------------
> x←2 2⍴'XX' '' '' ''
> s←(⍳2)∘.,⍳2
> ((∊2=⍴¨x)/,s)⌷x
> RANK ERROR
> ((∈2=⍴¨x)/,s)⌷x
> ^ ^
> ------ expected (no RANK ERROR, equivalent result as in previous 4x4
> matrix)
> XX
>
> Best Regards
> Hans-Peter
>
>
>
>
>
>
>
