stupid question (quick search did’t get me anything) how to I access to all function arguments from within the function? Basically looking for an equivalent of arguments object in javascript

@ag I think most likely reading this might help out : http://theburningmonk.com/2013/09/clojure-multi-arity-and-variadic-functions/

It's like doing argument destructuring in ES6, basically

Or this perhaps, http://clojure-doc.org/articles/language/functions.html#variadic-functions

@ag -- the vars are bound and without macros, you can't really do that. however, this may give you some direction: http://stackoverflow.com/questions/18797856/get-clojure-argument-list

Also there are no stupid questions ¯\(ツ)/¯

I have a question though. Functionally the following are the same, afaict, to get a value within a multi-level hashmap. Is either better, either idiomatically, functionally or performance-wise?

(-> msg :author :username)
(get-in msg [:author :username])

idiomatically, no. both are common.

functionally, I’d say basically no

(-> msg :author :username)

if that is important of course

boot.user=> (def m {:a {:b 34 :c {:d 42}}})
#'boot.user/m
boot.user=> (time (dotimes [_ 10000000] (-> m :a :c :d)))
"Elapsed time: 300.658938 msecs"
nil
boot.user=> (time (dotimes [_ 10000000] (get-in m [:a :c :d])))
"Elapsed time: 632.927241 msecs"
nil

I think there are two reasons for that - get-in reduces over the keys, invoking get on each key where as -> is compiled (via macroexpansion) into the direct series of lookups (:username (:author msg))

Thank you both for the thorough answer!

I shall be using -> from now on ^_^

@eslachance I think they both have a place 🙂

i.e. -> is a macro so it does not compose

and the second reason is that the -> will use (:author msg) vs (get msg :author) which is ever so slightly faster

agreed - get-in is great when you have some code that creates the lookup vector

Ah yes, doing it programmatically would need the vector

That's a good point

have do I get names of symbols (not values) that passed into function?

e.g. this (def foo “foo-val”) (name (quote foo)) works, right?

but this (map (fn [e] (name (quote e))) [foo bar]) wouldn’t

I guess I need to use some quoting trick

@ag you need a macro for this. when you call (my-func foo), foo is evaluated before being given to the function. the answer here should show you how to do it, in case you want to go this route. http://stackoverflow.com/questions/41482431/return-concatenated-params-in-clojure/41490018#41490018

@beppu awesome, I was looking at the generated API: http://clj-time.github.io/clj-time/doc/index.html and it does not have that method. Its a bit of a pity. Anyway, thats what I need, thanks

how can i remove all keys from a nested map? i.e. i want to strip away all keys == :db/id. I was thinking about postwalk, but have problem with coming up with how to tell it to remove the keys...

keys with values or just strip away the keys?

user=> (def x {:foo 1 :bar 2 :baz {:foo 1 :bar {:foo 2}}})
#'user/x
user=> (w/postwalk (fn [x] (if (:foo x) (dissoc x :foo) x))  x)
{:bar 2, :baz {:bar {}}}

@bronsa thx 🙂

i’m not sure i understand this:

`user=> (= (concat "Abc: " nil) "Abc: ") false`

how do i get that value to be true?

@pgmcgee concat returns a sequence, while "Abc: " is a string

either do (seq "Abc: ") or (apply str (concat "Abc: " nil))

oh, derp derp, it looks like i can also use (str “Abc:” nil) in this case

thanks so much!

@moxaj Thank you very much!!

Watching @plexus's latest episode on defrecords (https://lambdaisland.com/episodes/defrecord_defprotocol), it struck me that I've seen reader literals for defrecords used only rarely - in fact I forgot about their existence.

I mean something like this (from the transcript)

#ep24defrecord.core.Book{:title "Iain M. Banks"
                           :author "Consider Phlebas"
                           :year 1987}

What are those reader literals useful for?

serializing/deserializing record instances

ah because the records are automatically printed as literals

so basically you can easily serialize those literals to EDN or transit?

as a side node this works in a Clojure repl but not in lumo

(do (defrecord Foo [name age]) (cljs.reader/read-string (pr-str (->Foo "Joe" 92))))
Error: Could not find tag parser for cljs.user.Foo in ("inst" "uuid" "queue" "js")

no idea what lumo is

@anmonteiro's self-hosted clojurescript repl

https://github.com/anmonteiro/lumo

but I don't think cljs.reader implements type reader literals

so "type read literal" is the concept I was missing

if anyone's following along, here's the relevant section in the docs: https://clojure.org/reference/reader#_deftype_defrecord_and_constructor_calls_version_1_3_and_later

>>> Calls to Java class, deftype, and defrecord constructors can be called using their fully qualified class name preceded by # and followed by a vector

so this sounds as if this would work with regular Java classes too

#java.util.Date[0]
#inst "1970-01-01T00:00:00.000-00:00"

hmm... not seen the lambda island before.... is it worth it?

@pesterhazy it does, it's just that record use that as their default printer while other instances use #<Klass [email protected]>

@bronsa,

#java.util.Random[]
#object[java.util.Random 0x73600add "[email protected]"]

ah yeah, that's the new readable printer

so it's `#object[Klass hash "[email protected]"] ?

yeah

since 1.7 IIRC

I forgot about it because I monkey patched it to work as before in my local repls

too noisy :)

#object can't be read though (No reader function for tag object)

it's readable

what's readable as opposed to "can be read"?

if you define a tagged reader for object then supposedly you can make something useful out of it

it can't be read by default

but it's just a tagged reader away

ah I get the distinction

vs #<..>

which is unreadable

it's readable in the sense of potentiality, not an actuality

yeah

nice

it's kinda cool what you can uncover

btw the last item I think is just the toString representation of that instance

[class, hash, tostring]

@theblackbox to your question, I can wholeheartedly recommend lambda island, I've been a subscriber from the beginning

I've subscribed since the beginning and it's some of the best information you can find, for beginners to intermediates (with some advanced stuff sprinkled in). Disclaimer: I'm a friend so potentially biased

cool, good to know... I'll have a look see

Speaking of which, can anyone recommend a great book on Java?

Java is a pretty big topic, what are you looking into

I'd love to learn about advanced topics

intermediate to advanced I guess

memory model, concurrency, performance, GC

This is pretty much the final word on concurrency (and memory to some extent): https://www.amazon.com/Java-Concurrency-Practice-Brian-Goetz/dp/0321349601

Also any Java books on here would probably be good: https://www.amazon.com/gp/richpub/listmania/fullview/R3LG3ZBZS4GCTH

> Original Cotton Rope Hammock with Stand 😂

@tbaldridge good stuff, thanks

"The Java™ Virtual Machine Specification" ... not sure if I'm prepared to go there yet...

pesterhazy I found this book extremely helpful in my early days of programming: https://www.amazon.com/Effective-Java-2nd-Joshua-Bloch/dp/0321356683/

@schmee looks good

got all of the above. Used books are so affordable 🙂

hello guys, I'm struggling to pick my way through the various concurrency methods in an attempt to debug my code. I believe I'm in a race condition whereby a process is kicking off before a doseq has finished initialising various stuff and things so I would like to block until I know initialisation is complete. Where should I be looking? Futures look most likely?

Futures do allow you to block if that’s what you’re asking

what “process is kicking off” are you referring to?

so I'm using Clojure to implement Kafka Streams using interOp =/

I have my "topology" built by one process and my thinking is that the error I am currently getting means that I am calling "start()" before the topology has been initialised

this process that builds your topology — you are calling it how?

using "thread first"

sorry this is literally my first week with Clojure so I may not be using any of the words correctly 😉

@theblackbox Is it possible to post the relevant code portion?

so I'm trying to break out the "process" method so that I can throw as many predicates at it as I need and it should implement a stream for each

then it should start but I just get the error: "Exception in thread "StreamThread-1" java.lang.IllegalStateException: Consumer is not subscribed to any topics or assigned any partitions"

which is exactly what the process method should be doing

Got it. Admittedly, I'm not well versed in Kafka. But I'll take a look and see if anything is glaringly off.

cheers =D

from what i can tell, (doseq [predicate model/all] process) .. is not doing what you think it should be doing.

more than likely

as in when I use a single predicate - it works

you want:

(doseq [pred model/all]
  (process pred))

assuming model/all is a vector of predicates

me too

so yeah, if it is implemented like that using the single predicate, it works as anticipated

(doseq [predicate model/all] process) … evaluates to nothing useful. you’re not calling process like you seem to want to do. try the above.

@joshjones that seems to get me further, cheers

@theblackbox might want to use the .. macro instead of ->

(.. builder (stream input-topic) (map kvmapper) etc)

sorry what's the advantage there? or is it rtfm?

I think it's a readability thing.

cool, I'll look into it

From the doc: (.. System (getProperties) (get "os.name")) expands to: (. (. System (getProperties)) (get "os.name")) but is easier to write, read, and understand.

It may get a bit awkward with the length of your chain though.

hello guys. can someone enlighten me how to implement pretty standard pattern “worker pool with bounded task queue” in clojure? i’ve already got ztellman/manifold in my project and it will be pretty cool if it will fit.

that's pretty much exactly what core.async was designed to do

@nwalkr see: https://clojure.github.io/core.async/#clojure.core.async/pipeline

ok, i’ll take a look

@nwalkr also look at https://github.com/tolitius/lasync

@tolitius i’ll try it too. it seems to better fit my case out-of-box.

thanks.

Does lein test run tests in parallel by default?

https://github.com/weavejester/eftest does

But regular un-pluginned lein doesn't, right?

Trying to figure out why a test is failing, looks like things are running in parallel when we don't want them to.

not sure

try adding println at the beginning and end of tests to see if your hypothesis is right 🙂

lein test doesn't run in parallel by default

I also suspect it's not lein test

but that doesn't mean that your code's not running in the background because one of your tests kicked off some background process by accident

which is really easy to do if the code your testing has a lot of future or core.async code in it

If anyone is interested, I am trying initialize an open source project that facilitates building campaigns for public office in a transparent and less costly manner. There are more details in the #political-canvas channel

https://github.com/dmillett/political-canvas

howdy, I’m trying to get single-writer, multi-reader broadcast working in core.async; I tried using mult and tap, but that seems to turn the write side into a dropping buffer

don't think so, but the mult does't remember past values, so it's best to setup your taps/mult and then start putting in values

if you post some code, we can probably give you better advice

My writer code is almost literally (alt! [[ch value]] :sent (timeout to) :timeout), and right now the read side is (alt! ch ([v] v) (timeout to) :timeout)

what I was after is having multiple readers receive a value when sent, and to have the writer time out if no-one is reading

are pub and sub the things to use instead? I was going to study them next

@csm pub/sub is basically mult/tap but with topics. So you can dispatch to only certain channels based on the "topic" of each message

Where would be a good place to ask a question about clojure.java.jdbc?

here 🙂

Awesome. So here’s a problem I’m having. Running this:

(jdbc/query db [”SELECT my_table.* FROM my_table WHERE my_table.id = ?” id])

[{:id “someid”}]

What I would like is for it to return:

[{”my_table.id” “someid”}]

By default clojure.java.jdbc will “uniquefy” column names by appending “_N” to it, but I need to be able to prepend the table name to the column name

You may need to do that in sql.

not sure though.

usually best not use * at all

can't you use my_table.id AS my_table.id ?

Nah, that’s not valid SQL (the alias can’t have a period in it)

then use another character and replace? 🙂

it's hard to advise without knowing the use case

this may be to do with the JDBC driver as well, not clojure.java.jdbc per se

hihi. is it ok if I use one folder for cljc, clj and cljs?

@leov, yeah I don't see why not

i think you can give the jdbc stuff a function to run on your keys afterwards

https://clojure.github.io/java.jdbc/#clojure.java.jdbc/query

Usage: (query db sql-params)
       (query db sql-params opts)
Given a database connection and a vector containing SQL and optional parameters,
perform a simple database query. The options specify how to construct the result
set (and are also passed to prepare-statement as needed):
  :as-arrays? - return the results as a set of arrays, default false.
  :identifiers - applied to each column name in the result set, default lower-case
  :qualifier - optionally provides the namespace qualifier for identifiers
  :result-set-fn - applied to the entire result set, default doall / vec
      if :as-arrays? true, :result-set-fn will default to vec
      if :as-arrays? false, :result-set-fn will default to doall
  :row-fn - applied to each row as the result set is constructed, default identity
The second argument is a vector containing a SQL string or PreparedStatement, followed
by any parameters it needs.
See also prepare-statement for additional options.

specifically, the :identifiers

you can transform your column names however you like

You still wouldn’t know the table names though right?

his query has the table name in it

so presumably he knows it

if that’s the case can’t he just wrap the returned value?

(jdbc/query db ["SELECT my_table.* FROM my_table WHERE my_table.id = ?" id]) is his example, so prepend the known table name?

@nwjsmith did you see that documentation above?

I do, but I don’t think that will quite do it

I’m trying to write a function that takes any result set and transforms it into a normalized map structure in memory

It won’t have knowledge of the query, but it could have knowledge of the schema

I’m just surprised you can’t hook into the mapping of ResultSet -> seq of maps translation the way I want to

I essentially need to replace clojure.java.jdbc/result-set-seq with my own function

Does the java ResultSet return the table names?

indirectly, through the ResultSetMetaData https://docs.oracle.com/javase/7/docs/api/java/sql/ResultSetMetaData.html#getTableName(int)

you can hook into this with some of the lower level things in java.jdbc

clojure.java.jdbc/db-query-with-resultset

keep in mind transformed columns could possibly not have a table name

but like @dpsutton said -- you already know the table name a priori so you might as well use it

http://clojure-doc.org/articles/ecosystem/java_jdbc/using_ddl.html

You can probabluy use the metadata-result stuff?

So I think that clojure.java.jdbc/db-query-with-resultset

will do what I need

👍

thanks @7h3kk1d !

lol, I tried.

Hey all, so I have an array of objects, and they are a subset of the overall data, so I want to assign them their proper ids (shift by some offset). So, I assume the best way is to convert it into a map right? If so, I cannot seem to figure out how to convert an object from an array to a map.