Clojurians Log v2

boop.core> (def m {:username "sally" :profile {:name "Sally Clojurian" :address {:city "Austin" :state "TX"}}}

boop.core> (get-in m [:profile :address :state])
"TX"
boop.core> (:state (:address (:profile m)))
"TX"

You have everything right except 'adress'. It's likely, as @caumond says, the spelling.

All the books can be found here https://practicalli.github.io/

Ah I didn't see it as a title in that big cluster of graphics! Yes, here's the web apps link directly for @mircoporetti https://practicalli.github.io/clojure-webapps/

Great list. I’ll add “Joy of Clojure” - one of my favorites. https://www.manning.com/books/the-joy-of-clojure-second-edition

Keep asking questions, the Clojure community is one of the biggest advantages to learning Clojure https://practicalli.github.io/blog/posts/cloure-community-getting-help/

I think most languages are niche, until one of three things happen: * A killer app in the language blows up: for example Rails put Ruby on the map. Spark put Scala on the map. Numpy put Python on the map. * A large company becomes the shepherd of the language. Google most famously pushed Python and Go. Mozilla pushed Rust. Twitter pushed Scala. Microsoft pushed TypeScript. * A platform forces its use. Swift and ObjectiveC for Apple. C for linux. JavaScript for browsers. Java/Koltin for Android. C# for Windows and Unity.

All very good points. JavaScript is a particular good example. I don't think JavaScript was considered a serious language for a long time. Then jQuery came along (which still seems the most used library). Then more business got behind the language (even Microsoft) and of course every companies started using JavaScript for front end websites. Then node and mom came along... Although there are no guarantees that a language grows even with a 'killer app` or company support (i.e Dart from Google) or enviable platform changes. I am enjoying the diversity in languages we have experienced, even though it can add a burden to keep up. This diversity does encourage developers to thing differently about the software they write. This thinking is beneficial to Clojure as there are many subtle and important differences to discover

hello all, I am trying to make an app in cljs (with reagent) and it needs to interact with an API that's being hosted by a django backend. Is there a way to do the neat little development environment but somehow proxy all my requests to the django backend?

In this video, there's a quote from Rich Hickey and the presenter says that if there's one thing to be taken away from the talk, it should be that:

When you combine two pieces of data you get data. When you combine two machines you get trouble.

@dan.yb.tobias IIUC your django backend will be the only web server, right? then just tell your build tool to put the generated javascript somewhere that your django backend will know to serve

in principle it's not very different to working with a Clojure backend

doesnt figwheel or whatever spin up a little development server?

i dont want to have to compile every time i make a change etc

that's independent of the server that actually serves your main web page

https://figwheel.org/docs/your_own_server.html

ty

if you serve a html doc like the one described in that page from your web server, and if your web server serves the right file when asked for "/cljs-out/dev-main.js", you should be good

It comes from here: https://www.infoq.com/presentations/datomic-functional-database/

14:48

Hi everyone, I'm not sure where I should ask this but am trying to get the automated testing for clojurewerkz/money working again. Seems like at least for openjdk8 all tests pass but am encountering Tried to use insecure HTTP repository without TLS as shown here https://travis-ci.org/github/clojurewerkz/money/builds/753470823 Understand that this is the result of the snapshot deployment link for Nexus, but does anyone know what link i should update it to in the project.clj here - https://github.com/clojurewerkz/money/blob/master/project.clj#L19

Transcript is here if you prefer reading: https://github.com/matthiasn/talk-transcripts/blob/master/Hickey_Rich/FunctionalDatabase.md

@zackteo Change http:// to https://?

Hey guys and girls. New to Clojure, I'm learning by writing clojure to solve problems at http://projecteuler.net. I just solved #3 using Clojure and I'd like to get some advice and tips about how I could've written my code better. Here's a gist, lmk if the link works: https://gist.github.com/orpheus/e803cb8fb90f925fda90b07576273741 Also, there's an error I could use help debugging with the getFactor function, if passed 6857 (`getFactor 6857`) there's a stack overflow that I'm not sure how to go about solving.

@seancorfield thought about that. Guess I could give that a try. Was thinking if the link was meant to be inaccessible

@titanroark a quick thing to try would be to replace recursive calls to getFactor with recur: https://clojuredocs.org/clojure.core/recur

im a bit confused here, the doc linked doesnt say anything about it

do i need to change the port of the ring server to match as well?

also, to find factors you only need to check up to the square root... but now we're talking math. if your current primary focus is learning Clojure, http://projecteuler.net problems may have the focus too much on the math side?

I am so confused, am i running fig:build or what?

factorial is a different math concept. stick to the name "factor"

euccastro i think what you suggested before isnt quite right

lines 8-11 are the same as line 14, just specialized for the number 2. you could just call (getFactor x 2) at that point

your django server would replace the ring server. it doesn't matter what port it listens on, as long as you visit localhost:<that-port> in your browser

you can run any regular figwheel command

sorry, I had missed your reply; I've just replied in that thread

there is no html file produced though?

oof this is not working at all, i made an html and loaded the js there but

I did some Project Euler problems some time ago. You can find here: https://github.com/nakiya/euler/blob/master/euler.clj Just check for problem 3

@euccastro it looks like how the js is trying to load files is bad...it's using escape charatcers in strings?

Wow it did work. Guess i should have just tried that

https://github.com/jstaffans/django-cljs-loader i got it working using this

Not sure if this falls under beginners but has anyone encountered a situation where a new github pull request (that just changes the readme breaks the existing clojure tests? CI on github actions) https://github.com/zero-one-group/fxl/pull/31/checks?check_run_id=1666803003 It is very strange. Because when I try rerunning the tests on the previous commit there is no issue. But once I make a new pull request, even in the minimal example just amending the readme, breaks the CI

It's possible the tests are non-deterministic or something in the tests is not compatible with the GitHub Actions environment...

hmmm, my issue is that it would make sense that the github actions doing the test would have failed for the previous commits then or failed when I reran the actions but no issues

It is just generally quite an odd situation

Computers are deterministic by definition 🙂

My own theory is that the github secrets might be having issues when I am the one pushing the commit

Tho I am not sure if that theory checks out with the github actions history :thinking_face:

In which it should be intermittent and kicking off the process again should succeed...

Oh I mean the owner of the repo might be the only one able to access the secrets - I can't tell if that makes sense

I do know that this exists tho https://github.blog/2020-08-03-github-actions-improvements-for-fork-and-pull-request-workflows/

guess all I can do is sleep on it

Do not know what is the right channel for pedestral questions so I ask here I have this :

`(ns hello
  (:require [io.pedestral.http :as http]
            [io.pedestral.http.route :as route]))

(defn respond-hello [request]
  {:status 200 :body "Hello, world!"})

(respond-hello {})

deps.edn

{:deps
 {io.pedestal/pedestal.service {:mvn/version "0.5.7"}
  io.pedestal/pedestal.route   {:mvn/version "0.5.7"}
  io.pedestal/pedestal.jetty   {:mvn/version "0.5.7"}
  org.slf4j/slf4j-simple       {:mvn/version "1.7.28"}}
 :paths ["src"]}

but still I see this error :

; Execution error (FileNotFoundException) at hello/eval13893$loading (REPL:1).
; Could not locate io/pedestral/http__init.class, io/pedestral/http.clj or io/pedestral/http.cljc on classpath.

im following this tutorial now : http://pedestal.io/guides/hello-world

You have a typo, it's called pedestal, not pedestral. (Remove the r ) FYI there's #pedestal

thanks

What am I not getting?

(defn vec-of-14 [& seqs]
  (count seqs) ;=> 3
  (->> seqs
       (concat)
       (take 14)
       (vec)) ;=> Eval timed out!
  )

(comment
  (vec-of-14 (repeat 14 true)
             (repeat 1 false)
             (repeat true)) ;=> Eval timed out!
  
  (->> (concat (repeat 14 true)
               (repeat 1 false)
               (repeat true))
       (take 14)
       (vec)) ;=> [true true true true true true true true true true true true true true]
  )

You probably need to replace (concat) by (apply concat) in your method.

Oh, of course… Thanks!

See that I have to learn a lot more about clojure to make my own sites and really understand how things really work

interceptors are I think the hard part

You don't need interceptors to make a site

Start simple

Pedestal bundles a few things together

If you find it hard to grok you could go a step back. I think the absolute minimum for a beginner is something like compojure and figwheel but it's best to ask around

maybe I will or go back to duct or take a few steps more back. Before pedestal I did a few tutorials about duct and before that about ring.

what would be the pattern if I basically want to block until a certain condition is true? I'm waiting on a file and the last-modification time keeps changing and I want to wait until it doesn't change any longer for 100ms or something

Your deps.edn looks fine. But you have a misspelling in your require. "pedestral"

thanks

So what exactly are the differences between `(filter #(<= 5 %) coll)` and `(filter (partial <= 5) coll)` ? Is that second example using a "transducer?"

From reading the home page docs I don't think so. I should be creating an "xform." I get the feel that transducers are the current trend in performant, idiomatic clojure but I'm mostly confused about it all.

Any resources you recommend to start thinking in the transducer way of doing things? I feel like that could (should?) be the next step in upping my fp game.

What initially inspired this is this reddit thread: https://www.reddit.com/r/Clojure/comments/kr0j0f/newbie_question_on_idiomatic_iteration/ I want to start thinking like how SimonGray and joinr handled the problem

https://clojure.org/reference/transducers

neither of those are using transducers

I knew the first one wasn't, just wasn't sure if that second one was. I guess I don't really know what I'm asking besides how to start thinking in the higher abstraction, transducer sense of solving problems. Probably the only answer is experience.

read that guide and internalize what transducers are. i think its really important to remember they are functions that take reducing functions and return reducing functions

Will do. The embarrassing part is the first paragraph reiterates: Note: this reduced arity is not currying or partial application. And here I am asking if that partial produced a transducer. lol

no worries. they can be a bit difficult to grok them. follow that guide and build up your intuition and also keep that definition in mind: a function from a reducing function to a reducing function

And are they more performant to use in most cases? I know that's probably to general to ask though. But is it something you usually try and reach for when writing idiomatic clojure?

at the top is a link for "good use cases for transducers" which has a section about performance. nothing more authoritative or thoughtful than that section

Fair enough! Thanks

again stuck how I have to deal with this :

; Implement update-in.

(def my-update-in [m k f]
  
  )

Sounds like a good plan, yes. If by "put the whole m again together" you are thinking "create a new map in N steps if the original map m has N key/value pairs", then there are faster ways, e.g. "create a new map from m by replacing only the value of key k with the new value"

hmmm, then I have to think again

that could work but on some way I have to check where I want to check things

You could get a correct solution by creating a new map in N steps.

n steps ?

It probably just would not be among the fastest solutions.

Your original statement said "and put the whole m again together".

There are multiple things you might have meant when you said that, and I don't know what you were thinking.

that is not the most important I think. This is a challenge from the brave book and I think the purpose is to understand how things work

One thing you might have meant was "for each of the N keys in the original input map, add that key to a new map to be returned"

not find the fastest one

Lets say we have this map : (def users [{:name "James" :age 26}  {:name "John" :age 43}])

That is a vector containing two maps.

oke

another one then : {:name "James" ;age 26}

That is one map 🙂

now I want to change the age to 36 because I made a typo

sure

then my idea was to use for example `get to get the age out of it

sounds good

make a new age with 36

yep

and then I have to make a new map with the name part and the new age part

yes

or if I want to change the name a new map with the new name and the old age part

but I think I do not see how the pieces comes together

My only point was that there is more than one way to do that last step, and I was making a comment about efficiency of a couple of different ways. If you aren't worried about efficiency at all, then by all means ignore that comment.

I think I have to store the new part somewhere with a let ?

I wouldn't say "have to", but if it helps you write a solution, then let is perfectly reasonable.

(def my-update-in [m k f] (let [new_value (f (get m k)]      )

If you want to use let, then taking your breaking the function up in to steps, you can first get the current value in the map for key k, and bind that result to some new name in the let, e.g. cur-val or whatever you would like to call it.

Sure, what you do there is doing your first two steps in one line, which is also good.

oke , if I go to fast , say it please

I will. Don't worry, I'm following so far 🙂

then I have to make a new map. I think I need to use assoc

but how do I now from the old map what is then the old part ?

and what is the new part?

So assoc always returns a new map, leaving the original one unchanged. The new map is almost the same as the input map, except for the value associated with the key you give to assoc, which in the returned map will be the value you give to assoc

yep, so there im stuck

user=> (assoc {:a 1 :b 2 :c 3} :b -2)
{:a 1, :b -2, :c 3}

oke, so I do not have to know

I didn't tell assoc to do anything with keys :a or :c, and the returned map still has them, associated with their original values.

That is what assoc is intended to do.

oke, did not realize that

Has the book you are following described assoc before this exercise, I hope?

so I can do just this :

(def my-update-in [m k f]
   (let [new_value (f (get m k)]
   (assoc m new_value) 
  )

yep, I did

but I think I misunderstood it as just for adding things

Your call to assoc doesn't have the correct number of parameters.

(def my-update-in [m k f]
   (let [new_value (f (get m k)]
   (assoc m k new_value) 
  )

that looks better

Looks like a good time to try out your new function with some sample input parameters in a REPL and test if it works as you hope.

there is still a error somehow

What editor are you using?

calva is yelling me that the closing ] is on the wrong place

I use vs code with calva

Most editors have the option to show you when parens, square brackets, or curly braces are matching or not. You definitely want to get familiar with how to understand what it is telling you there.

Anyone using cider with Doom emacs? Trying to figure out the most appropriate way to evaluate the form under the cursor

e.g. many editors will show you the matching paren, bracket, or brace when your cursor is on or just after one of them, highlighting it visually so you can see which one matches the one where the cursor is.

yep, I was missing a )

Trying to write a Lisp-like language without such editor support is going to slow you down tremendously.

There is a lot of fancy 'structural editing' stuff you may have heard about that some people love, but I've never used that myself. But I do really, really miss it if I ever use an editor that doesn't show me matching parens and brackets.

still somewhere a error :

; Execution error (ClassCastException) at chapter5/my-update-in (form-init95645667318631163.clj:51).
; class java.lang.Long cannot be cast to class clojure.lang.IFn (java.lang.Long is in module java.base of loader 'bootstrap'; clojure.lang.IFn is in unnamed module of loader 'app')

(defn my-update-in [m k f]
  (let [new_value (f (get m k))]
    (assoc m k new_value)))

(my-update-in p :age 35)

Is there anything else in the same file with that function? Looking at the function definition, nothing obviously wrong jumps out at me.

What is the 3rd parameter of my-update-in supposed to be?

Im my first idea it was a function but later it changed to the new value

What does my-update-in do with the 3rd parameter?

it uses it to make the new map with a updated key

so it is not a function

The only place the 3rd parameter is used in my-update-in is this expression (f (get m k))

So it is trying to call f as a function.

yep, I see it , thinking how tto solve it

That is what the exercise asked you to write, I think, yes?

Then when you are trying to use my-update-in, you gave it the number 35 as the 3rd parameter, and my-update-in tried calling the number 35 as if it were a function.

So it appears to me that your my-update-in is doing what the exercise says it should, but your attempt to call the function is passing a 3rd parameter that it was never intended to handle.

i.e. you are calling my-update-in incorrectly. You should try calling it where the 3rd parameter is some function.

If you want a quick try, replace 35 with inc, which is a function.

I'm using it with lispy[ville] and lispy-eval keybinding (iianm I had to add com.cemerick/pomegranate and com.billpiel/sayid to my clojure deps for it to work)

yep, that works

Does it make sense why you got the error when you passed in 35?

it makes sense

Do you know how to write a Clojure function that takes a number and adds 10 to it?

yep

Good. If you define that function, you could pass add-ten as the 3rd parameter to my-update-in

oke, so there is no way I could make it work without a function ?

only given the right value

You could write a different function than my-update-in

that takes a new value for the 3rd parameter instead of a function.

If you did, I would call it my-assoc 🙂

I was thinking to change my_update_in

If you change my-update-in to take a value instead of a function as the third parameter, you can certainly do that. Note that it would not be one that solves the exercise you were working on.

?? the only that the challenge said it to implement update-in

Yes, I understand that.

YOu asked if you could change my-update-in to behave differently. You can do that if you want.

I am only saying that such a changed my-update-in does not solve the challenge.

oke, I see it, update-in takes always a function

as the 3rd parameter, yes.

so I wanted to make another function

And I encourage you to experiment to your heart's content 🙂

thanks for thinking with me and teaching me some things

no problem

clojure is a totally other beast then when I was learning ruby or haskell

but I love the way repl helps me to see the results when I try something

You could write my-update-in in Haskell in about the same number of characters, I suspect. Clojure and Haskell share that they deal mostly with immutable values, but Haskell has far fewer "escape hatches" than Clojure does for mutating state.

And Clojure doesn't restrict you to writing code that passes Haskell's type checker.

and clojure is much faster then a haskell programm is my feeling

and I played with some web stuff and clojure has much more choices then haskell

in haskell you have only yesod and servant as big players

I don't know a lot about Haskell performance to compare it to Clojure -- they both are not necessarily the language you want to use to get the best performance for heavy numerical computation, for example.

clojure has for back-end duct, pedestral en some more

but both are perfectly fine for calling out to libraries written in other languages for doing that kind of thing.

but much more to do for your own

dinner time

and tomorrow or later time to study the next chapter macro's 🙂

Apart from that and cider-eval-… bindings as well as cider-pprint-eval-last-sexp-to-comment I so far have found cider debugger with enlighten-mode quite useful. Seeing how flexible this all is I'd be also interested in seeing what others are using for quick & easy inspection of forms & values 🙂

I moved away from lispy because of le-clojure.clj. I use cider-eval-defun-at-point and cider-eval-last-sexp ... If I ever need to be more granular, I use er/expand-region and then cider-eval-region ... all these are bound to things like , e d , e e etc

Use while

These are the same

The only difference is that #(<= 5 %) creates the predicate function at read time, and (partial <= 5) creates the predicate function at run-time

The docstring you're reading is referring to the arity (number of arguments) of filter

Both your example call filter with 2 arguments though, so in your case you are using the 2-ary version of filter in both case

The doc says return a transducer when no collection is provided, and in both cases you provide [pred coll]

it really doesnt look like cljs works well with a django backend 😛

i am really struggling to get it working

The backend and front end are orthogonal. If you return Json the backend technology is irrelevant for cljs. Do you have a more specific problem you can share?

i think it's more about the project.clj file, im doing a figwheel-main reagent template and I guess i need to hook it up so that the project itself isnt necessarily in the django folders but the produced html/js/css ends up in the correct places

there's just like 2 million directories specified in this thing

looks like the configuration doesnt like relative paths either

damn

can you share your project.clj?

@b Are you're referring to the part of lispy provided by le-clojure.el ? Does it have some flaws one should be aware of?

(defproject livematch "0.1.0-SNAPSHOT"
  :description "FIXME: write this!"
  :url "<http://example.com/FIXME>"
  :license {:name "Eclipse Public License"
            :url "<http://www.eclipse.org/legal/epl-v10.html>"}

  :min-lein-version "2.7.1"

  :dependencies [[org.clojure/clojure "1.10.0"]
                 [org.clojure/clojurescript "1.10.773"]
                 [reagent "0.10.0" ]]

  :source-paths ["src"]

  :aliases {"fig"       ["run" "-m" "figwheel.main"]
            "fig:build" ["run" "-m" "figwheel.main" "-b" "dev" "-r"]
            "fig:min"   ["run" "-m" "figwheel.main" "-O" "advanced" "-bo" "dev"]
            "fig:test"  ["run" "-m" "figwheel.main" "-co" "test.cljs.edn" "-m" "livematch.test-runner"]}

  :profiles {:dev {:dependencies [[com.bhauman/figwheel-main "0.2.12"]]
                   
                   :resource-paths ["target"]
                   ;; need to add the compiled assets to the :clean-targets
                   :clean-targets ^{:protect false} ["target"]}})

the main issue im having i think is the dev-main.js can't find any of the deps it needs for some reason

if i look at dev-main.js i have lines like this:

document.write('<script src="..\..\paladeenz\website\static\livematch\public\cljs-out\dev/cljs_deps.js"></script>');

notice the escape characters..

it's just copy and pasting the target directory..but remember the src is in a different folder than my django app...so when the js gets put into the django app it needs to look in the correct place for the deps but isnt

what’s dev-main.js?

where is this file? and what’s suppose to do?

this is the output js from building...i fixed it by editing all of the urls the issue is that if recompiled it will overwrite and i have to change it over and over

ohh ok…

the target-dir i set is so that it can create dev-main.js in the correct spot, but dev-main itself seems to be using the target-dir also to find the dependencies which completely breaks it

so i have a figwheel-main.edn file which says

:target-dir "../../paladeenz/website/static/livematch"

which puts dev-main.js where it needs to be..it's just now dev-main.js is trying to load files from target-dir + other stuff, which is wrong

https://figwheel.org/config-options#target-dir idk if there's anything in here to fix that

it’s long time since i use fighweel.main but let me see if i can find something

im going to try adding in output-to and output-dir

NVM it didnt like those keywords haha

it's like this stuff is designed to be as cryptic as possible

did you try maybe asking in the #figwheel-main channel, i think somebody with more experience than me will be able to help you (and wife is screaming at me now 8:30 pm now :D)

ok ill ask there, ty

What is here wrong ?

(defn my-update-in [m ks f args]
  (assoc m ks (apply f (get m ks) args)))

(my-update-in p [:1] :age inc)

; Execution error (IllegalArgumentException) at chapter5/my-update-in (form-init95645667318631163.clj:51).
; Don't know how to create ISeq from: clojure.core$inc

you specify f before args, then :age before inc

also (get m ks) is weird - it's as if you expect the whole seq of keys to be one key

neither :age nor inc are collections, so neither works as a last arg to apply

I will sleep about this one :

(defn my-update-in [m ks f & args]
  (let [[[k & ks] ks  (assoc m k (apply f (get m ks) args))]])

; Syntax error macroexpanding clojure.core/let at (chapter5.clj:51:3).
; [[[k & ks] ks (assoc m k (apply f (get m k) args))]] - failed: even-number-of-forms? at: [:bindings] spec: :clojure.core.specs.alpha/bindings

that last call needs to be outside the [] binding block

the syntax of let is (let [binding value ... ...] body) - it's usually a logic error not to provide a body, and it's a compilation error to have a binding without a value (or visa versa)

When first learning, I would recommend avoiding the use of destructuring, until and unless you get a version of a function working without it.

Destructuring can be useful, but it is never necessary.

hmm

you mean like this :

(defn my-update-in [m ks f & args]
  (let [[k & ks] ks]
  (assoc m k (apply f (get m ks) args))))



(my-update-in p [:1] assoc :name "James" :age 35)

that does this :

{:name "James", :age 26, :1 {:name "James", :age 35}}

and not updating the value

oke

hmm, still no luck

(defn my-update-in [m ks f & args]  
  (assoc m ks (apply f (get m ks) args)))



(my-update-in p [:0] assoc :name "James" :age 35)

{:name "James", :age 26, [:0] {:name "James", :age 35}}

maybe the apply is wrong here

@andy.fingerhut

Im not getting what you want to achieve there. Update-in work with nested maps. And ks is supposed to be a list a key of your map. What do you want to specify with :1 ?

but I do not see any better one

It is almost always possible to take a Clojure function and evaluate it step by step in a REPL session.

Do you know which sub-expression of your my-update-in above must be evaluated first?

I would say get m ks

yes, good.

And what parameters it should be passed in your example call, and what value it will return?

(If you are not sure, that is where the REPL can be handy)

i think the old value is taken where then the function is applied to

It isn't clear to me what the value of p is in your call above

sorry, I forget to paste that

(def p {:name "James" :age 26})

we solved that one earlier but I try to make it work that it also worked with (def users [{:name "James" :age 26} {:name "John" :age 43}])

So what does (get m ks) return?

in the example of p it will return 26

Do you prefer to try to figure out why your current function is behaving like it is? Or do you prefer to start with the input/output behavior of the function you want to write, and try to write that?

Now, try eval'ing this in a REPL: (get p [:0]

and see what it returns.

I think it schould be (get p p: [:0])

that returns nil

so not good

Does it make sense why it returns nil?

yep, I think there is no key named [:0]

correct.

So we can keep looking at what happens with your current function, but maybe you might prefer starting with explaining what function behavior you are hoping to write, because I don't know what that is yet.

You say you want a function that takes a vector of maps like this: (def users [{:name "James" :age 26} {:name "John" :age 43}]) and does something with it, but I'm not sure what you want the behavior to be.

I was hoping to get first a record from a map so in this example {:name "James" :age 26}

So in Clojure terms, that would be getting the first element from a vector or sequence.

and then I schould find a way to update what I want to update

The value of users is not a map, but a vector of two elements (each of those elements is a map)

yep, or the second if I want to update the second entry

So try to explain what the parameters of the function you want are, and see if you can give one or two examples of what you want the return value to be for those example inputs, along with a sentence explaining what the function ought to do.

oke, in the example of p. Lets say I want to update the age to 45

Right ?

then m is still p so {:name "James" :age 26}

So p is a single Clojure map, right, meaning this one: (def p {:name "James" :age 26})

yep

I work from that example first

later we can do more entries

So you wrote my-update-in earlier that could take m, a key k, and a function f, and do inc or add-ten, or whatever function you want, to the current value associated with that key.

That was working.

So now you are saying you want a function that can take a vector of maps, and do something with that.

yep

My first comment would be: you could write a single function that does both, but maybe given that the first argument is a map in one case, but a vector of maps in the other, maybe a different function might be better.

oke

to be sure we are on the same page

i.e. try writing a function that only works if the first argument is a vector of maps, without trying to serve both purposes.

I want to end at a function that can do this :

(def m {:1 {:value 0, :active false}, :2 {:value 0, :active false}})

(update-in m [:1] assoc :value 1 :active true)

outcome : 

{:1 {:value 1, :active true}, :2 {:value 0, :active false}}

Once you can see both functions on their own, and get them working, then maybe go back and see if it makes sense to try to write some combination of them that can do both behaviors. But I suspect it will only be more confusing to try to write that combined-behavior function first.

oke, no problem to take one of more steps back

So in this example call (update-in m [:1] assoc :value 1 :active true) you have a vector containing :1. Do you want this proposed update-in function to be able to take any number of keys in that vector, 1 or more? Or do you want it to only work for 1 element in that vector?

If you only want it to work for 1, then why put it in a vector?

maybe better to make it work with 1 element first

and then look what needs to be changed for multiple entries in the vector

I guess perhaps from the name update-in, you might be thinking along the lines of how get-in works in Clojure?

Are you familiar with get-in?

yep, that one is explained in this chapter as Whereas the function get-in` lets you look up values in nested maps, `

So if your update-in is intended to be similar to get-in, in that it can take a vector of keys, and update a value in a nested map, then it makes sense that it should be a vector. In fact, your update-in would likely want to use get-in to get the current value.

moment

it does not work as expected :

(def m {:1 {:value 0, :active false}, :2 {:value 0, :active false}})

(get-in m [:1][:active] )

{:value 0, :active false}

How many keys to you give to get-in /

?

Oh, I see what you did. That is not how you give a sequence of multiple keys to get-in. Try this: (get-in m [:1 :active])

1 sequence of keys, not multiple separate sequences of keys.

yep, that works

get-in takes an optional 3rd parameter, that if you give one, and if the sequence of keys you give are not found in the collection, it returns the 3rd parameter to indicate "not found", instead of nil

In this call: (get-in m [:1][:active]) there is a key :1, so the 3rd parameter is not used.

yep, I see it : (get-in m [:3 :active] "not-found")

gives not found

yep. as it is documented to do.

(once the documentation makes sense, that is 🙂

yep, and I tried to see if I really understand things

experimentation at the REPL with different variations of an idea are definitely good ways to test your understanding.

yep, I will , can we talk tomorrow about this further

its late here, 22:36

sure. I will warn you that I expect update-in will be a bit more code to write than my-update you wrote earlier, but doable with patience.

and I will experiment if I can make it work on a different keyword, so on value and active

It is definitely a good idea to have a good understanding of what Clojure built-in functions like get and get-in do, especially when relying on them to write your own functions.

and then try to adapt it to make it also work on the different entries so the keyword and the number of entries (1 or 2)

so first understand that

and then look how I could change a value and make a new map

thanks for the lessons

you are the first one which do not give a answer but let me think about a solution

for now GN

It usually takes longer to ask questions that let the learner think about a solution :-)

My hope is that you will learn questions to ask yourself, and habits of thought, that let you more often find the answers yourself.

and got the code where the number and the keyword is a variable

(defn get-on-keyword [m keyword number]
  (get-in m  [number keyword] "not-found"))

and now really sleep

Maybe a idea to try now t change the get-in into update-in ? after some sleep naturlly