Clojurians Log v2

I've not used it personally, but this sounds like the kind of problem https://github.com/redplanetlabs/specter might be useful for.

using clojure's walk fn you can check if a value is a MapEntry with the given key and update its value

walk + "a pretty large data structure" is not always the best combination if you care about performance

I thought pre/post/walk will feed my function the key and the value separately. Do those also pass in MapEntrys?

@vale yep

@lasse.olavi.maatta what else could I do?

I'm not saying there's some single obvious alternative, it depends on the situation if you can avoid walking. For example, in one situation I had code walking over several large but structurally static(!) datastructures. I split the walking into separate "find out what paths to update" and "actually update the value behind each path" steps and got like 50-200x speed up because the paths were static and I calculated them just once.

I usually try to think "why don't I know the exact paths to the values I need to alter". Could the producer of the data provide some extra information which I could use later to construct the paths, so that I don't have to resort to walking? But of course, this isn't always possible.

I see your point. update-in just gets so clumsy when there are vectors involved somewhere in the structure

Also note there's nothing that prevents nesting for or doseq expressions; for simple cases you can even use the built-in iteration over multiple variables:

(for [a [1 2 3]
      b [4 5 6]]
  [a b])
;; => ([1 4] [1 5] [1 6] [2 4] [2 5] [2 6] [3 4] [3 5] [3 6])
(doseq [x [1 2 3]]
  (doseq [y [4 5 6]]
    (println [x y])))
;; => nil
;; but prints:
[1 4]
[1 5]
[1 6]
[2 4]
[2 5]
[2 6]
[3 4]
[3 5]
[3 6]

(loop [outer 3]
  (loop [inner outer]
    (when (pos? inner)
      (println inner)
      (recur (dec inner))))
  (when (pos? outer)
    (recur (dec outer))))
;; => nil
;; but prints:
3
2
1
2
1
1

Just noting that doseq supports the same syntax as for, so no need for nested doseq here - it can be a single doseq.

Yes; sorry if it wasn't clear, I just meant to show that there's nothing that prevents nesting.

I saw the documentation of clojure.core/for and that is what I am using...

is this wrong?

there is a for list comprehension that lazily builds a list. Which requires a consumer of that list

there is no such thing as a for loop in clojure core though

clojure.core/for
([seq-exprs body-expr])
Macro
  List comprehension. Takes a vector of one or more
   binding-form/collection-expr pairs, each followed by zero or more
   modifiers, and yields a lazy sequence of evaluations of expr.

This is from the example blurb

ClojureDocs Examples
 via cider-nrepl

Example 1
To Rich Comment | To Output

;; prepare a seq of the even values 
;; from the first six multiples of three
(for [x [0 1 2 3 4 5]
      :let [y (* x 3)]
      :when (even? y)]
  y)
;;=> (0 6 12)

yes. there the repl’s printer is forcing the realization of the list. What is forcing the realization in your example?

ok. Then do you have a way of nesting loops one inside the other?

functions usually

ok

I'll give it a try

Your example will work if you replace for with doseq. And then you also won't need that do.

@p-himik it did indeed

thanks!