as long as oK is the only one that uses unicode that's fine, it's just surprising that there's only 1 (or maybe not considering the k aesthetic of keeping simplicity)
also, technically k supports any single-byte encoding, not necessarily ascii
ascii is only 0..127
i wonder if the people who were designing ascii at the time realised how much impact it would have on the world.. it's hard to imagine a single-byte (or multibyte) encoding used now that's not ascii-compatible.
Background
K functions have a feature called projection, which is essentially partial application of values to a function. The syntax for projections is a natural extension of the regular function call syntax:
f[1;2;3] / call a ternary function f with three arguments 1, 2, 3
f[1;2;] / the ...
@ngn can you explain a bit how &/~=/2#+-2!<'+,/ works? I think I understand what it does, but not why applying it gives the right answer. Grade the coordinates, flip, div 2, flip again, take the first two....
cmc: given a list of even-length int lists, pad them with 0s so they are centred, e.g. (1 2;3 4 5 6 7 8;9 1 2 3) -> (0 0 1 2 0 0;3 4 5 6 7 8;0 9 1 2 3 0)
two ideas: 1) `k pretty-print should do that table formatting thing argued about here a while back and 2) it would be much shorter with ⎕FLIP:`center; (presumably adding trailing whitespace as well, which would be OK)
I just want to write ? instead of distinct, is that so much to ask :P
they could've at least gone with uniq...
it's totally weird, they have long names for the primitives (the things you use a lot and can remember easily) and then short cryptic names for all the other utility stuff (the things you don't use a lot and probably can't remember easily)
@rak1507 yeh I think / and `\` in scripts are the multi-line comment syntax
@rak1507 hmm. I can see two usecases for heavy usage of SQL-y things; data science-y type stuff (basically whatever folks use numpy/pandas for) and things that shuffle a lot of data around
the moment you start writing actual q code anyway any illusion of being a 'normal' language is quickly shattered when you have to do $[] for if statements and whatever, idk why learning # means count and / means fold was too hard for people
@rak1507 I think I read/watched something that mentioned that the dyadic versions were kept since they're more familiar (presumably meaning the arithmetic ones like + or *, even though things like x#y and x_y would still be unusual coming from other languages)
@rak1507 depending on the K I would include things like if[c;expr1;expr2;...] plus all the defined/keyworded functions. and the SQL stuff itself is a nice stepping stone (given that a "select col1 * col2 from table" "vectorizes" the way it does in array languages)
Background
K functions have a feature called projection, which is essentially partial application of values to a function. The syntax for projections is a natural extension of the regular function call syntax:
f[1;2;3] / call a ternary function f with three arguments 1, 2, 3
f[1;2;] / the ...
