The APL Orchard

apl.chat ― Learn, teach, ask, code, golf, & discuss usage. See ...
Jul 1 16:44
@Silas Good to know, thanks.
Jul 1 15:02
OK, thanks.
Jul 1 14:35
Anyone know how to install a user command on a Mac?
May 22 21:54
@B.Wilson Ties survive workspace loading, so I wonder if way back in the day of limited workspace size, a constant literal tie number was useful. If you had the tie number in a variable you would lose when loading a different workspace.
Mar 19 19:02
Signing off...
Mar 19 19:02
Will check it out, thanks.
Mar 19 18:56
I assume the under operator will help with transposing twice?
Mar 19 18:55
Don't be, I'm a rank beginner, pun intended
Mar 19 18:54
Ok. I feel better now.
Mar 19 18:53
I think that is my problem. If I want to apply a non-scalar user-defined function to the columns of m, I can't do that directly with rank without transposing first. Is that correct?
Mar 19 18:48
Yes, not saying good, trying to apply rank'
Mar 19 18:48
Trying to apply user defined function f to column of M
Mar 19 18:47
hold, not sure my question is well formed...
Mar 19 18:46
What about doing ⍉↑v f¨↓⍉m
Mar 19 18:45
This is how I understood it, so the reference to column on APL Cart confused me.
Mar 19 18:45
"Effectively, we are taking each scalar from v and pairing up with each row from m."
Mar 19 18:40
Please do...
Mar 19 18:40
Ah....
Mar 19 18:38
What am I not getting?
Mar 19 18:38
Yes, but there are only 3 column vectors on the right
Mar 19 18:37
I'm sure now I don't understand rank
Mar 19 18:34
Is that right, or it swapped for the definition of the corresponding one for rows?
Mar 19 18:33
is "Apply f between vector Mv and each column of Nm"
Mar 19 18:32
APL Cart says that Mv(f⍤0 1)Nm
Feb 17 18:13
@Adám Nice, will check it out.
Feb 16 13:32
@Adám Constructing the attribute column for ⎕XML for HTML tables. Pairing up attribute names and values.
Feb 14 00:09
Signing off for today, thanks again for your help.
Feb 14 00:08
Can you guess the application?
Feb 14 00:06
I'm liking e3
Feb 14 00:06
Values are all char vecs, not scalars, in my particular case.
Feb 14 00:05
 cmpx e1 e2 e3 e4
  ⊂[1 0](d⍴(×/1↓d)/j),[¯.5]k → 2.9E¯4 |    0% ⎕⎕⎕⎕⎕⎕
  ⍉⊂⍤2⊢j,⍤0⍤1⍉k              → 3.8E¯4 |  +32% ⎕⎕⎕⎕⎕⎕⎕⎕
  ⊂[0 3]j,⍤0⍤0 2⊢k           → 2.6E¯4 |   -9% ⎕⎕⎕⎕⎕
  (↑(⊂¨j),¨⊢)¨,⌿⊂¨⊂¨k        → 1.9E¯3 | +575% ⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕
Feb 13 23:25
Ok, now you just caused me another day of study!
Feb 13 23:22
often because they don't come up, but often because its easier to cheat and use each
Feb 13 23:22
Yes, I really need to study them more. It's remarkable how much coding one can do and avoid them
Feb 13 23:20
Ok, noted. I'll take that under advisement.
Feb 13 23:20
I'm pretty sure I never would have figured that out!
Feb 13 23:19
Cool, thanks so much!
Feb 13 23:19
Both do same!
Feb 13 23:18
and it seems to do what I want...
Feb 13 23:18
I just tried ⊂[0 3] j,⍤0⍤0 2⊢k
Feb 13 23:13
Cool! Am actual doing a ⊂[1 0] on my result, so maybe I can get rid of the transpose...
Feb 13 23:12
Woah... is that rank twice or an atop? I'm going to have to spend a day studying it!
Feb 13 23:05
Funny it actually runs with the extra letter
Feb 13 23:04
Yes, sorry
Feb 13 23:02
using the rank operator to avoid all of the reshaping of j?
Feb 13 23:02
        (d⍴(×/1↓d)/j),[¯.5]k
Feb 13 23:01
is there a way to do this:
Feb 13 23:00
   d←5 4 3
   k←d⍴⍳×/d
   j←'abcdef'
Feb 13 23:00
Given:
Feb 4 21:42
In production to would default to ⍺ ;)