I'd like to help with my 2 cents.
Given your comparison with R, sum, and mean are expected to play with a
vector rather than log and sin are expected to play with single numbers.
Then, the expected behavior for numerics types in Raku still the same as in
R. The difference is only that the functions in Raku are designed to one
and only one single function.
#RAKU
> my @nums = 0,1,2,3,4,5,6,7,8,9;
[0 1 2 3 4 5 6 7 8 9]
> @nums.sum
45
> sum(@nums)
45
#R
> library(tidyverse) #import pipe %>%
> nums <- 0:9
> nums
[1] 0 1 2 3 4 5 6 7 8 9
> sum(nums)
[1] 45
> nums %>% sum()
[1] 45
> nums %>% log
[1] -Inf 0.0000000 0.6931472 1.0986123 1.3862944 1.6094379 1.7917595
[8] 1.9459101 2.0794415 2.1972246
> log(nums)
[1] -Inf 0.0000000 0.6931472 1.0986123 1.3862944 1.6094379 1.7917595
[8] 1.9459101 2.0794415 2.1972246
In this point, the unique weirdness I'd like to understand is why in Raku
`@nums.log == 2.302585092994046e0`. I don't understand where this value
comes from.
In the case of strings, pay attention to the size of the array. The infix
`>>` act as expected keeping the vector with the same size, which is the
default behavior of R, since it is functional and vectorization is arguably
its philosophy which handles very well data analysis which is its purpose.
> @monsters>>.split(" ").elems
5
> @monsters.split(" ").elems
7
> my @monsters = << blob 'king kong' mothera fingfangfoom 'foo bar'>>;
[blob king kong mothera fingfangfoom foo bar]
> @monsters.elems
5
> @monsters.split(" ")
(blob king kong mothera fingfangfoom foo bar)
> @monsters.split(" ").elems
7
> @monsters>>.split(" ")
[(blob) (king kong) (mothera) (fingfangfoom) (foo bar)]
> @monsters>>.split(" ").elems
5
The point is, I don't know why Raku is designed the way it is. But is
easier to implement in Raku the use of functional and vectorized paradigms
than in Python which requires an external lib (e.g. Pandas or numpy).
Of course, it might be better if Raku has docs explain the paradigms and
philosophy behind the scenes and very didactic and well-explained docs as
Perl5. Also, coping R, might be too inviting to learners and seniors if we
had online handbooks, wiki-style docs, and cheat sheets as we have in R.
best
On Wed, Oct 14, 2020 at 3:28 AM William Michels via perl6-users <
perl6-us...@perl.org> wrote:
> On Mon, Oct 12, 2020 at 10:02 AM Larry Wall <la...@wall.org> wrote:
> >
> > On Mon, Oct 12, 2020 at 01:14:09PM -0300, Aureliano Guedes wrote:
> > : > This seems pretty convenient and intuitive. At least, it is possible
> > : > to mimic that behavior in Raku:
> > : >
> > : > List.^find_method('split').wrap: { $^a.map: *.split($^b) }
> > : > List.^find_method('sin').wrap: *.map: *.sin;
> > : >
> > : This is like overwrite the function?
> > : Might be better just implement a new function, I mean, a new verb as is
> > : called in R.
> >
> > In Raku these functions already have names, if you count metaoperators
> as a fancy
> > way to name anonymous functions:
> >
> > say <a,b c,d>».split(',');
> > say (0, (π/2, 3 * π/2))».sin;
> >
> > ((a b) (c d))
> > (0 (1 -1))
> >
> > As shown by the ».sin example, unary hypers will distribute over
> > multi-dimensional structures for you, just as an APL or R programmer
> > would expect. But that behavior is not intuitively obvious to everyone,
> > so the vector-processing paradigm is not the default. (But we make it
> > really easy to get to when you want it, as you can see. And arguably
> > the explicit presence of » or >> makes your intent clearer to the naïve
> > reader, who at least has something to look up or ask about if they don't
> > understand it.)
> >
> > Larry
>
> Hi Larry!
>
> First of all, let me thank you for designing the Raku (née Perl6)
> programming language in the first place. I've really enjoyed learning
> a different set of programming paradigms than I was previously
> accustomed to.
>
> With regards to the present topic, what initially 'disunited' me was
> calling split() on an a array and ending up with joined elements. It
> becomes most apparent when one tries to split() on a character that
> isn't present in the array in the first place--in that case, all array
> elements are joined into a single string. It has been explained to me
> that arrays coercible to strings and called with a string-function
> *are* auto-joined with a single space as separator, but it would be
> nice to control this behavior somewhat, e.g. for generating a CSV line
> (join on commas instead of spaces).
>
> However. I guess the real question in my mind is how to predict which
> function "verbs" will work on array "nouns" in plural form (I'm daring
> here to discuss linguistics with a linguist). Certainly we know nouns
> that remain invariant from singular to plural (sheep, fish, deer,
> salmon, aircraft and other -craft, etc.). These nouns exist and don't
> seem to be going away anytime soon. I can say "Eels fill my
> hovercraft," and that one statement applies to both singular and
> plural hovercraft. But swapping the statement around to the more
> familiar, we clearly see the verb lets us know whether hovercraft is
> used in the sigular or plural: "My hovercraft is/are full of eels."
>
> Thinking of Raku arrays as invariant nouns, I might wonder which
> "verb" functions denote acting on a 'singular' array (acting on the
> array as a whole) versus acting on a 'plural' array (i.e. acting
> element-by-element). Without consulting the Docs, in regards to
> functions acting wholly or primarily on strings, I might guess that
> the 'grep()' verb acts on a 'singular' array. From my previous posts
> you can gather I guessed that split() acts element-by-element on a
> 'plural' array (along with many other string functions). Without
> consulting the Docs, in regards to numeric-functions I might guess
> that the sum() verb and any 'mean()' verb would act on a 'singular'
> array, and I'd guess that most other numeric functions like log() and
> sin() would act element-by-element on a 'plural' array. This could be
> completely off-base, and a consequence of sum() or mean() being
> 'many-to-one' functions, while log() and sin() etc. have a 1-to-1
> input/output correspondence.
>
> Now maybe this is a road that has been trodden before, with Perl. Or
> maybe it's my own linguistic preconceptions getting in the way. But I
> really do hope to understand why the Raku language is designed the way
> it is designed.
>
> Thanks, Bill.
>
> > #REPL code below:
> Nil
> > my @monsters = << blob 'king kong' mothera fingfangfoom >>;
> [blob king kong mothera fingfangfoom]
> > dd @monsters
> Array @monsters = ["blob", "king kong", "mothera", "fingfangfoom"]
> Nil
> > @monsters.grep(/" "/);
> (king kong)
> > @monsters>>.grep(/" "/);
> (() (king kong) () ())
> > dd @monsters.split(" ");
> ("blob", "king", "kong", "mothera", "fingfangfoom").Seq
> Nil
> > dd @monsters>>.split(" ");
> Array element = [("blob",).Seq, ("king", "kong").Seq,
> ("mothera",).Seq, ("fingfangfoom",).Seq]
> Nil
> > my @nums = 0,1,2,3,4,5,6,7,8,9;
> [0 1 2 3 4 5 6 7 8 9]
> > dd @nums
> Array @nums = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
> Nil
> > dd @nums.sum
> 45
> Nil
> > dd @nums>>.sum
> (0, 1, 2, 3, 4, 5, 6, 7, 8, 9)
> Nil
> > dd @nums.log
> 2.302585092994046e0
> Nil
> > dd @nums>>.log
> Array element = [-Inf, 0e0, 0.6931471805599453e0,
> 1.0986122886681098e0, 1.3862943611198906e0, 1.6094379124341003e0,
> 1.791759469228055e0, 1.9459101490553132e0, 2.0794415416798357e0,
> 2.1972245773362196e0]
> Nil
> >
>
>
> https://dictionary.cambridge.org/us/grammar/british-grammar/nouns-form?q=Forming+the+plural+of+nouns
> https://www.lexico.com/grammar/matching-verbs-to-collective-nouns
>
> https://dictionary.cambridge.org/us/grammar/british-grammar/nouns-singular-and-plural?q=Collective+nouns+%28group+words%29
> https://www.lexico.com/grammar/plural-nouns-treated-as-singular
> https://omniglot.com/language/phrases/hovercraft.htm
> https://www.theguardian.com/notesandqueries/query/0,,-197456,00.html
>
--
Aureliano Guedes
skype: aureliano.guedes
contato: (11) 94292-6110
whatsapp +5511942926110
