« first day (2911 days earlier)      last day (33 days later) » 

Adám
Adám
05:42
@Richard Both do work, but won't accomplish the task here.
 
10 hours later…
Silas
Silas
15:39
@Adám huh?! monadic ⍪ doesn't seem to accept an axis specifier...
Adám
Adám
16:15
@Silas No, only dyadic. What would it even do monadically?
 
1 hour later…
Silas
Silas
17:17
@Adám indeed - was confused by previous message which seemed to say both axis and rank apply to ⍪ in both monadic and dyadic forms
 
6 hours later…
Silas
Silas
22:58
@Adám {⊃⍪/⊂⍤2⊢2 1 3⍉⍵}x obviously :)
Adám
Adám
@Silas Sure, but what does it do?
Silas
Silas
interlace alternate rows
taking from each matrix - hence thought should be doable with transpose
Adám
Adám
Transpose cannot combine axes and also preserve all elements; it takes diagonals.
Silas
Silas
Feel there should be a nicer way to combine the layers after the transpose - although guess that's what ⍪ is for 
          2 1 3⍉x←2 3 4⍴⍳99
 1  2  3  4
13 14 15 16

 5  6  7  8
17 18 19 20

 9 10 11 12
21 22 23 24
and then you want to stitch each block together was my thought
@Adám not using the transpose to combine - wanted a nicer way to stitch the layers
Adám
Adám
,[⍳2]2 1 3∘⍉?
I've been promoting (no pun intended) ,[⍳2] as a primitive, Demote.
Silas
Silas
23:10
thanks
that's what I was after - don't know that axis
Annoyingly BQN's monadic ⍉ rolls the other way - thought that would remove the 3d restriction
Adám
Adám
Right, axis operator is, off-kilter.
@Silas 1. You could use ⍉⁼ 2. Still wouldn't help, as it only gives you access to 2 3 1⍉ and 3 1 2⍉ but not 2 1 3⍉
Silas
Silas
oh, yeah - permutation not rotation
Adám
Adám
And dyadic ⍉ is exactly "permute axes".
Silas
Silas
@Adám writing out the axis, this kinda makes sense - combine across these axes 
,[1 2]2 1 3∘⍉ ⍝ join 1 and 2 axes together of the transpose
Silas
Silas
23:46
Think general case is: 
{,[⍳¯1+≢⍴⍵](1⌽⌽⍳≢⍴⍵)⍉⍵}

« first day (2911 days earlier)      last day (33 days later) »