I’ve never seen a deref used like this before. What does it mean?

@(def sample-data
   (->> (for [color colors size sizes type types]
          {:inv/color color
           :inv/size size
           :inv/type type})
        (map-indexed
         (fn [idx map]
           (assoc map :inv/sku (str "SKU-" idx))))
        vec))

the def returns a var, and deref, when applied to a var, "returns its current state"- so say the docs.

so it will return the results of- constructing a list of :inv/color :inv/size :inv/type maps, passed through map-indexed, and then turned into a vec

@jonahbenton Ah, I see, thank you. So instead of returning #'some.ns/sample-data it’ll return the the contents of it.

It did do that, but I didn’t quite understand what I was looking at.

yeah, definitely an opaque formulation for clj, though it feels idiomatic for something, but I can't place it. maybe scala is where i remember seeing phrases like that. I would guess it was written by someone working in a functional-but-not-immutable mindset.

I encountered it in the Datomic tutorial: https://docs.datomic.com/cloud/tutorial/assertion.html#sec-5

I guess it was written like that in order to show what’s contained in sample-data.

Any way to trick gen-class in running at runtime?

Is there a purpose difference between funcool/struct and clojure/spec ?

Does anyone have a solution to the way cljfmt indents cond (I prefer the tests and expressions to be indented differently)? I have an elisp rule for cider, but I'm trying to create the same rule for those that don't use emacs.

This is the formatting I find much easier to read:

(cond
  t1
    e1
  t2
    e2)

@didibus (compile ...) of a namespace with gen-class doesn't produce .class files?

does anyone know of any really good functional cookbooks or resources for working with really multidimensional nested data?

I'm basically forming compound keys for vertical and horizontal subslices of multidimensional data and I feel like a fool. I can't do my normal pipeline approach because I have to map over the thing I'm mapping over the thing I'm mapping over 😩

I'm trying to string together some crazy apply map comp stuff, and I get the distinct feeling this is like a rubiks cube that I don't know the patterns for

multidimensional nested data as in arrays of arrays of arrays etc? or nested maps?

its mixed. arrays of maps of arrays of arrays of maps of maps

are you aware of Specter? sounds like it might help with that https://github.com/nathanmarz/specter

in this particular instance, it's because I'm working with hiccup. But in general I've hit these kinds of problems before and I feel like I need to find the spirit of Peter Norvig

oooh, I was not aware of Specter. this looks interesting

+1 for specter. it is a bit tricky to learn but totally worth it

as for your question, this is great but you may find it slow going at the beginning, depending on your background: http://www.ccs.neu.edu/home/matthias/HtDP2e/

oooh, that's awesome! a library and a "hot to get gud" class? That's outstanding!!

skim part five to get a feel about whether this is what you're looking for: http://www.ccs.neu.edu/home/matthias/HtDP2e/part_five.html

the "Design of Computer Programs" that Peter Norvig teaches on Udacity is one of the best things I've ever done for my professional career

same here 🙂

(def data-1 `{(a 1) :c})
(type (first (keys data-1)))
;; => clojure.lang.Cons
(def data-2 {(list `a 1) :c})
(type (first (keys data-2)))
;; => clojure.lang.PersistentList

how do I convert nested data with clojure.lang.Cons in it (like data-1) to normal one (like data-2)?

if your code relies on the difference between cons and list, that is usually a problem with your code

What is your usecase? Things like Cons and PersistentList should usually be an implementation detail.

If you need a predicate to check for a list, I did a take on this some time ago: http://bytopia.org/2016/03/08/what-is-a-list/

@noisesmith @alexyakushev I need to validate some data with clojure.spec, but it says the data in data-1 is invalid while data-2 is ok

Is that a spec you wrote or have control over?

does your spec use list? if so, don't

list? is pretty much never useful

I wrote the spec

Then you've probably used a check for list? there somewhere where you shoud've used seq?

really, I'd expect s/cat or s/tuple etc. rather than either...

It could be (s/coll-of any? :kind list?)

hmm, looks like the problem is not with the spec I wrote, but with the spec from a library I use which is a bit complex

@noisesmith @alexyakushev I'd rather stick to a solution to the question I originally asked: how to convert such nested data

update-in with list*

but it's a sign of a bad spec, it shouldn't care about the difference between a cons and a list

it can be in arbitrary shape

then clojure.walk/post-walk with seq? and list*

so no update-in

You can use clojure.walk/postwalk then.

it's still a lot of effort to accomodate a bad design

okay, I'll try

thanks a lot

Where can I find the method that turns java.util.Date into "2018-01-01T00:00:00.000-00:00" string?

it's defined with the reader tags if you mean the one that has #inst right before it

Does clojure.java.jdbc support executing multiple queries in the same round-trip?

@arohner same round trip? You mean like a transaction?

I’d like to send two different queries at the same time, and get both result sets back in one network round trip

I guess the same transaction, sure

hmm.. i mean you could do something async and a connection pool

why do you want to do that? maybe it'll help me understand....

(require '[clojure.spec.alpha  :as s]) 
(s/valid? int?     1) ;=> true
(s/valid? #(= 1 %) 1) ;=> true
(s/exercise int?   1) ;=> ([0 0])
;; ???
(s/exercise #(= 1 %) 1) ;; Spec assertion failed.  Spec: nil  Value: nil

@joefromct let’s say querying 2 different foreign keys from the same id

ok... should you be just... joining them in the db?

hey @arohner that's usually a wire protocol capability of the underlying database. you can, for instance, send multiple queries over the wire using the mysql protocol, but it's often disabled because it can be used for sql injection

when you brew install clojure it installs java for you right?

not in my experience

I don't think (not sure) postgres lets you fire multiple quiries delimited by ; if i remember correctly... sorta feels like bad form anyway (with params epsecially...?)

but if you have a table with an id and then two columns that are fks into their own lookup tables, you can get the data from the lookup tables with a single statement/join

and then I have to un-join the result on my end

I’m not proposing concat’ing strings

oh ok, so you want your results in two mutually exclusive seq's?

right

ok i might be missing it but whats wrong with this? probably not at the same time right... ?

(let  [seq-1 (j/query "select from table_one")
        seq-2 (j/query "select from fk_table") ]
...
)

yeah, though I guess I can try futureing them

once you have a connection pool, the overhead per-query is small. there's only so much you can fit into a packet.

I’m more worried about latency of waiting for query 1 to finish before issuing query 2

yeah unless the db is super slow... maybe table_one is a view/aggregate over 15 million rows

i'd put in in an async thread

yeah, agree ^^- all drivers will allow for true parallel communication

j/query is greedy too, so i've been using a lot of j/reducible-query lately, but if your results fit into memory no worries

think you need to provide a generator

a raw predicate function like #(= 1 %) doesn't have a generator attached, and spec needs a generator to exercise the spec

ah ok thats why it validates but i can't build me any data

ok

i saw that in the guide thank you

spec internally has a list of a lot of predicates from clojure.core that it implicitly has a generator for, like clojure.core/int?

yes

https://github.com/clojure/spec.alpha/blob/master/src/main/clojure/clojure/spec/gen/alpha.clj#L131-L188

@joefromct what version are you running? I see:

(s/exercise #(= 1 %) 1)
ExceptionInfo Unable to construct gen at

which is more obvious

[org.clojure/clojure "1.9.0"]

maybe it has to do with my repl

yeah it was dumping ^ to standard out i didn't see it

thanks