Presenting: polyfill

cool

not a fan of magic constants for reshape but the rest should definitely all be in the language

@rak1507 Do you have a better suggestion for auto-shape?

Hm, actually, I might change those "magic" values to be ¯1 ¯2 ¯3 rather than ¯0.5 ¯1 ¯1.5

Done, and added ¯4 to use fills. Now just like BQN's but using negative values instead of functions.

Got my book today :) Diving right in.

Why does ⍴ of a scalar (I think that's the right term) return nothing?

Oh I see. It has a rank of 0

They look the same, but they are different

1 is a scalar, 1⍴1 is a 1-item vector

      1 ≡ 1⍴1
0

1 2 3 is already a vector of shape 3

so reshaping to shape 3 has no effect whatsoever :)

Because 1 on its own is a scalar, whereas "1 2 3" can't possible be a scalar because it isn't 1 number

if you know python, it's sort of like (1) being a number vs (1, 2, 3) being a tuple (but only vaguely analogous)

Ok think I'm getting the idea

it wouldn't really make sense for 1 to be actually [1] under the hood (that would cause a lot of confusion and be annoying to work with), but it doesn't make any sense for 1 2 3 to be just a number so the only interpretation is [1, 2, 3]

Oh that makes a lot of sense

The default is to just be a number but multiple numbers can't be a single number

yep, that's pretty much how I understand it (probably worth confirming from someone who knows APL though - I can hardly say i understand any more than vaguely its train structure :P)

Thanks. I got it now

@hyper-neutrino It's a good intuitive explanation. A technical explanation is that multiple "things" (single numbers or arrays) put together get "stranded", which means they form a new vector containing them.

what's the difference between an array and a vector (in APL's context)?

A vector ≡ a 1D array ≡ a rank-one array

oh

wait, you can write multiple arrays together? i was under the impression that that wasn't really a thing you could do

like, to get a matrix you had to reshape or smth

Matrix does require a reshape or a mix. Stranding multiple vectors results in a vector of enclosed vectors

that's... a bit confusing. anyway, IIRC i was told there's a difference between a matrix and just sticking vectors in a vector, so now's a good time to ask - what is the difference exactly?

So the two common ways to write a 3x3 matrix containing 1..9 are:

      3 3⍴1 2 3 4 5 6 7 8 9
1 2 3
4 5 6
7 8 9

      ↑(1 2 3)(4 5 6)(7 8 9)
1 2 3
4 5 6
7 8 9

(reshape 1-9 as a 3x3, and mix a vector of enclosed vectors?)

The biggest difference is that a vector of vectors can be jagged (contain vectors of different lengths), but a matrix can't

@hyper-neutrino Yes

@Bubbler ah, that makes sense, thanks

(also from testing i notice reshaping insufficient values repeats elements and mixing jagged elements fills with 0s. which both make sense)

And they have different rank, shape, and depth, which heavily affects how all the structural functions work

,Y also doesn't seem to work on a vector of enclosed vectors

Because it only flattens the outermost array structure.

a matrix has depth 1?

that both makes sense and is sort of confusing

A "flat" matrix has depth 1. Of course a matrix (or any shaped array) can also contain enclosed arrays as items.

ah. okay. that makes sense

so basically, an array can have any number of dimensions but will still have depth 1

and if it nests enclosed arrays then it has higher depth

but you can have a matrix of enclosed matrices with depth 2 and dimensions 2 and 2?

Sure.

      2 2⍴⊂2 2⍴2
┌───┬───┐
│2 2│2 2│
│2 2│2 2│
├───┼───┤
│2 2│2 2│
│2 2│2 2│
└───┴───┘

and from what it looks like, does ⊂ "enclose" (like create a singleton array holding just that element)?

Correct.

cool. that... actually makes a lot of sense now. thanks :)

@mo523 if you use ≢ here instead of ⍴ it will work, ≢ of a scalar is 1 as well

@user8773069 @user334436 Hi John Doe and Stilvoid. If you want to participate here, please email access@apl.chat

@WongJiaHau Just don't expect functions to take functions as arguments (not that it is actually often needed), because only K and BQN can do that. See discussion of the subject here.

@ArpitSaxena you could try viewing it with ⎕SE.Dyalog.Utils.repObj

0⍴⊂'' probably

Yeah that was it. Thanks for that method, that will come in handy

ah yep, that. No monospace made me think there was a '|' character in the middle

yea I copied it to check

I couldn't figure out how to monospace here

as you type multiline, a "fixed font" button shows up on the right side of the chat box

you can also prepend four spaces to each line

┌⊖────┐
│ ┌⊖┐ │
│ │ │ │
│ └─┘ │
└∊────┘

(but you can't have a multiline message that's partly monospace and partly not)

Ohh

⊖ seems to be wider than other characters at least in the font on my system right now

so unfortunately it looks misaligned on my screen, lol. might try to set my font to something that supports more characters correctly

@hyper-neutrino And do that automagically by pressing CTRL+K

ah right, that too

never actually bothered trying that in chat before

Hey @user8773069 , if you would like to chat here, please email us at access@apl.chat !

(So that we can give you access aplwiki.com/wiki/APL_Orchard#Access)

CMC: codeforces.com/contest/1538/problem/C in under O(n^2) time

oh this is interesting

my thoughts; might be spoiler

right, that's what I thought at the time but doing that in APL will not be very elegant

@hyper-neutrino yeah, I think it's a spoiler

I will try it in Python first and then see if I can figure it out in Jelly and then APL (with my very limited knowledge of APL :P)

python has tio.run/##SypKzMxLK03O/v8/KbM4Nbnk/38A (spoiler) which makes it straightforward

oh yeah that thing

@rak1507 oooooooooooh I didn't know about that!

I did it in python during the actual contest

I think there is a pretty nice nlogn solution

I don't really do CF anymore

@dzaima does it use tio.run/##K85JLM5ILf7/PzOvJLWoLDFHITMvJbXi/38A

@rak1507 yep

:D

I think I have a solution too

Me too, I'll just do whatever he's ↑ doing

lol

and i seem to have an actuall working & executable (BQN) solution

wonder if you could persuade codeforces to add bqn lol

@dzaima and made an APL port, and a somewhat golfed 30 char BQN one

(oops: found bug)

(oh, mine just takes as a left arg [l,r] and a as the right arg. Didn't bother doing the whole fancy I/O)

wait nvm extended and unicode do different things with ⍸ for reasons

@dzaima yea fair enough I just did the io to 'verify'

as a function it'd be tio.run/##K85JLM5ILf7/…

I assume you did something similar

my APL solution

⎕IO←1?

yeah

need to figure out that ⌊¨ thing

@rak1507 tiny simplification based on mine

not sure I understand how the two are equivalent

@rak1507 something like this

ah, cool, thank you

that's smart although I think my original solution for me anyway is easier to understand

@rak1507 if i bothered to, it'd probably be easier for me to understand too. i barely understand what i did :p

number of pairs <= r minus number of pairs <= l-1, so the inner function calculates the number of pairs bit

my thought process

nice