👋 ☕ ☀️

Morning

Good Morning!

Morning!

Good morning!

morning

Goeiemorgen!

good morning

Working on a Bauhaus-inspired, mildly skeuomorphic aesthetic today for one of my frontends

måning

Trying to be faster, I discovered this little gem https://clojurians.slack.com/archives/C03S1KBA2/p1649401401958449

I really dislike these things that seem to work, until they don’t

Say you have two lists (of data retreived from a database) and you want one list. Which is more idomatic to use - concat, conj, merge....?

there won't be duplicates, as data is different

into ?

interesting, I've also discovered (mapcat identity...)

into works well too

moring... and TGIF again!!!

Morning

Good morning

I like into and cat

(into [] cat [vec-1 ,,, vec-n])

I like into too

@minimal so... you're into into. :drum_with_drumsticks:

i’m into into too @borkdude

concat works … the into [] is only needed if you want to force the resulting collection to be a vector. No?

I would also choose concat. Like @raymcdermott, I only use into when I need to enforce a collection type (or when reifying output from a transducer).

The implementation of the laziness of concat was problematic, can’t remember if it ever got fixed so I tend to stick with into

laziness never gets fixed

I also use concat, but some awareness around the pitfalls may be good: https://stuartsierra.com/2015/04/26/clojure-donts-concat

That’s the article I had in mind

BTW a really nice idiom is (into (empty coll) (f coll)) .

bah - if concat has been cancelled, I’m still unreconstructed

#cancelConcat

or perhaps f can be a transducer (if it has multiple map/filter, etc)

In Clojure we're doing quite a lot of work that is probably done by compilers in other more sophisticated languages ;)

(inspired by the latest #defnpodcast with @ben.sless;))

It’s the FFS pattern … for f’s sake use a transducer

I don’t use transducers or even parallelisation nearly enough.

if you still want to be lazy

(sequence cat [[0 1 2] [3 4 5]])
;; (0 1 2 3 4 5)

yeah if you still want concat/mapcat in the mix you can do something over the top with transducers (into [] (mapcat identity) [[1] [2] [3]]) [1 2 3]

aaand we're back to (into [] cat [[1] [2] [3]])

I'm a bit confused about what I've done wrong here:

(sequence cat [0 1 2] [3 4 5])
Execution error (ArityException) at user/eval15916 (REPL:68).
Wrong number of args (3) passed to: clojure.core/cat/fn--8851

as sequence has & colls as one of its arities

now I am confused

(sequence (mapcat identity) [0 1 2] [3 4 5] [6 7 8])
Execution error (ArityException) at user/eval15920 (REPL:74).
Wrong number of args (3) passed to: clojure.core/identity

sequence is like map, so (sequence ... xs xs2 xs3) = (map ... xs xs2 xs3)

so your fn will be called with as many args as you have sequence arguments minus the xf

ah, the problem is what I'm passing to mapcat and cat

(sequence (map (fn [x y z] [x y z])) [0 1 2] [3 4 5] [6 7 8])
;; ([0 3 6] [1 4 7] [2 5 8])

I was a bit surprised that (sequence cat ...) didn't do that

https://clojuredocs.org/clojure.core/cat

I think you are basically asking for (sequence cat [xs1 xs2 xs3])

it's not shocking to me, it translates nicely from (into cat ...)

sequence map with multiple collections is doing what I expect it to (getting the first then the second then the third from each collection)

then (sequence (comp (map (fn [x y z] [x y z])) cat) [0 1 2] [3 4 5] [6 7 8])

ok, this makes sense to me now at least

(sequence (comp
                 (map (fn [x y z] [x y z]))
                 cat)
                [0 1 2] [3 4 5] [6 7 8])
;; (0 3 6 1 4 7 2 5 8)

Good morning! I’ve recently come to realise that it can be hard to talk about the code in Clojure, as I’m pairing with someone essentially completely new to it. What seems simple to me can be troublesome to explain to them.

I mean, it’s probably the same for all programming languages. Having the same basic structure in most code, that is, (function-call arg1 arg2), helps a little bit, but there are so many exceptions to that.

(I’ve “taught” Clojure to people before, but this is a special case)

There is nothing wrong with concat. Contrary to into or anything with a transducer concat won't allocate a new collection/sequence. It's just a cheap lazy seq wrapper. And you can reduce that cheaply too. Just don't use it to iteratively stack up collections. apply concat is fine, reduce concat is not.

Welcome to my ted talk "have you considered not using a transducer?"