Contrived example to illustrate:
(def nest1 {:a {:b {:c "foo"}}})
(def nest2 {:d {:e "bar"}})
If I wanted to conj these nests at arbitrary levels, I could explicitly do this:
(conj (-> nest1 :a :b) (-> nest2 :d)) ; yields {:e "bar", :c "foo"}
(conj (-> nest1 :a) (-> nest2 :d)) ; yields {:e "bar", :b {:c "foo"}}
But what if I wanted to create a function that would accept the "depth" of nest1 and nest2 as parameters?
; Does not work, but shows what I am trying to do
(defn join-nests-by-paths [nest1-path nest2-path]
(conj (-> nest1 nest1-path) (-> nest2 nest2-path))
And I might try to call it like this:
; Does not work
(join-nests-by-paths '(:a :b) '(:d))
This doesn't work. I can't simply pass each "path" as a list to the function (or maybe I can, but need to work with it differently in the function).
Any thoughts? TIA...
Sean
Use get-in:
(defn join-by-paths [path1 path2]
(conj (get-in nest1 path1) (get-in nest2 path2)))
user> (join-by-paths [:a :b] [:d])
{:e "bar", :c "foo"}
user> (join-by-paths [:a] [:d])
{:e "bar", :b {:c "foo"}}
Your version is actually doing something like this:
user> (macroexpand '(-> nest1 (:a :b)))
(:a nest1 :b)
which doesn't work, as you said.
get-in has friends assoc-in and update-in, all for working with nested maps of maps. There's a dissoc-in somewhere in clojure.conrtrib.
(In Clojure it's more idiomatic to use vectors instead of quoted lists when you're passing around sequential groups of things.)
Related
I am looking to transform a clojure tree structure into a map with its dependencies
For example, an input like:
[{:value "A"}
[{:value "B"}
[{:value "C"} {:value "D"}]
[{:value "E"} [{:value "F"}]]]]
equivalent to:
:A
:B
:C
:D
:E
:F
output:
{:A [:B :E] :B [:C :D] :C [] :D [] :E [:F] :F}
I have taken a look at tree-seq and zippers but can't figure it out!
Here's a way to build up the desired map while using a zipper to traverse the tree. First let's simplify the input tree to match your output format (maps of :value strings → keywords):
(def tree
[{:value "A"}
[{:value "B"} [{:value "C"} {:value "D"}]
{:value "E"} [{:value "F"}]]])
(def simpler-tree
(clojure.walk/postwalk
#(if (map? %) (keyword (:value %)) %)
tree))
;; [:A [:B [:C :D] :E [:F]]]
Then you can traverse the tree with loop/recur and clojure.zip/next, using two loop bindings: the current position in tree, and the map being built.
(loop [loc (z/vector-zip simpler-tree)
deps {}]
(if (z/end? loc)
deps ;; return map when end is reached
(recur
(z/next loc) ;; advance through tree
(if (z/branch? loc)
;; for (non-root) branches, add top-level key with direct descendants
(if-let [parent (some-> (z/prev loc) z/node)]
(assoc deps parent (filterv keyword? (z/children loc)))
deps)
;; otherwise add top-level key with no direct descendants
(assoc deps (z/node loc) [])))))
=> {:A [:B :E], :B [:C :D], :C [], :D [], :E [:F], :F []}
This is easy to do using the tupelo.forest library. I reformatted your source data to make it fit into the Hiccup syntax:
(dotest
(let [relationhip-data-hiccup [:A
[:B
[:C]
[:D]]
[:E
[:F]]]
expected-result {:A [:B :E]
:B [:C :D]
:C []
:D []
:E [:F]
:F []} ]
(with-debug-hid
(with-forest (new-forest)
(let [root-hid (tf/add-tree-hiccup relationhip-data-hiccup)
result (apply glue (sorted-map)
(forv [hid (all-hids)]
(let [parent-tag (grab :tag (hid->node hid))
kid-tags (forv [kid-hid (hid->kids hid)]
(let [kid-tag (grab :tag (hid->node kid-hid))]
kid-tag))]
{parent-tag kid-tags})))]
(is= (format-paths (find-paths root-hid [:A]))
[[{:tag :A}
[{:tag :B} [{:tag :C}] [{:tag :D}]]
[{:tag :E} [{:tag :F}]]]])
(is= result expected-result ))))))
API docs are here. The project README (in progress) is here. A video from the 2017 Clojure Conj is here.
You can see the above live code in the project repo.
user=> (into {} '((:a :b) (:c :d)))
Throws: ClassCastException clojure.lang.Keyword cannot be cast to java.util.Map$Entry clojure.lang.ATransientMap.conj (ATransientMap.java:44).
Whereas:
user=> (into {} (list [:a :b] [:c :d]))
Is fine. It's a strange difference, since many times other functions return lists when the thing they had to begin with was a vector:
user=> (into {} (partition 2 (interleave [:a :b] [:c :d])))
Will throw, because it partition 2 ...) results in ((:a :c) (:b :d)). So it's pretty annoying. You basically have to memorize both the return types of methods and the specific behaviors of functions like into, or you have to just let stuff blow up and fix it as you find it with stuff like (into {} (map vec (partition 2 (interleave [:a :b] [:c :d])))).
Is there a specific reason why into doesn't like the pairs as lists?
The reason is as you state, only a vector pairs can be used to build maps. I don't know of a practical reason why this limitation exists. But there are also several other methods for constructing hash-maps. If you find yourself using partition, perhaps the answer is to use an alternate construction method.
If you have parallel sequences of keys and values:
(zipmap [:a :c] [:b :d])
If you have all the items in a flat sequence:
(apply hash-map [:a :b :c :d])
Building a map from a sequence:
(into {} (for [[k v] xs]
[k (transform v)]))
I never realized this wouldn't work! Don't forget:
(apply hash-map (interleave [:a :b] [:c :d]))
;=> {:b :d, :a :c}
since hash-map implicitly creates pairs from the scalar args:
(hash-map :a :c :b :d)
;=> {:b :d, :a :c}
you don't really need the (partition 2...) which is the source of the problem.
I have this function:
(defn dissoc-all [m kv]
(let [[k & ks] kv]
(dissoc m k ks)))
Where m is the map and kv is the vector of keys. I use it like this:
(dissoc-all {:a 1 :b 2} [:a :b])
=>{:b 2}
This is not what I've expected. ks has :b but I don't know why it is not being use by dissoc. Anyone can help me with this?
Edit: Added question is that why is this not triggering the 3rd overload of dissoc, which is dissoc [map key & ks]?
Changed name from dissoc-in to dissoc-all as noisesmith have said, -in is not a proper name for this and I agree.
This won't work because ks is a collection of all the elements in kv after the first. So instead of :b it is [:b].
Instead, you can just use apply:
(defn dissoc-in [m vs]
(apply dissoc m vs))
Also, dissoc-in is an odd name for this function, because the standard functions with -in in the name all do nested access, and this does not use the keys to do any nested access of the map.
Why not something like this?
(defn dissoc-all [m ks]
(apply dissoc m ks))
(dissoc-all {:a 1 :b 2} [:a :b])
=> {}
The reason the third overlod of dissoc is not getting called is because it does not expect a collection of keys like [:a :b] - it expects just the keys.
For example:
(dissoc {:a "a" :b "b" :c "c" :d "d"} :a :b :c)
=> {:d "d"}
Further to noisesmith's answer:
You're being confused by the overloads/arities of dissoc, which have this simple effect:
[m & ks]
"Returns a new map of the same (hashed/sorted) type,
that does not contain a mapping for any of ks. "
The explicit arities for no keys and one key are for performance. Many clojure functions are so organised, and the documentation follows the organisation, not the underlying idea.
Now, the action of
(dissoc-all {:a 1 :b 2} [:a :b])
;{:b 2}
is to bind
k to :a
ks to [:b]
Note the latter. The example removes the :a but fails to remove the [:b], which isn't there.
You can use apply to crack open ks:
(defn dissoc-all [m kk]
(let [[k & ks] kk]
(apply dissoc m k ks)))
(dissoc-all {:a 1 :b 2} [:a :b])
;{}
... or, better, do as #noisesmith does and short-circuit the destructuring, using apply at once.
I'm new to clojure and I've been staring at this for some time, I'm sure there's something basic I just don't see. I want to conj two sets, but they're nested, example:
(def foo {:b #{:test}})
(def bar {:a {:b #{:ab}} :c :d})
I tried:
=>(update-in bar [:a :b] conj (:b foo) )
{:a {:b #{#{:test} :ab}}, :c :d}
I guess that makes sense, but what I wanted was {:a {:b #{:test :ab}}, :c :d}
I just can't seem how to get either #{:test} out of the set to conj it, or to properly access :b as a set given the update-in syntax.
Any help is enormously appreciated.
You need to use into instead of conj:
(update-in bar [:a :b] into (:b foo))
;= {:a {:b #{:test :ab}}, :c :d}
I have a map in Clojure something like this:
(def stuff #{
{:a "help" :b "goodbye"}
{:c "help2" :b "goodbye"}
{:a "steve" :b "goodbye"}
{:c "hello2" :b "sue"}
})
: and I want to provide a search so that:
(search stuff "help")
: would return :
#{
{:a "help" :b "goodbye"}
{:c "help2" :b "goodbye"}
}
: What is the simplest way to do this?
user=> (defn search [s q] (select #(some (partial re-find (re-pattern q)) (vals %)) s))
#'user/search
user=> (search stuff "help")
#{{:a "help", :b "goodbye"} {:c "help2", :b "goodbye"}}
This does the trick.
Full text search is a different topic, but if you can live with regexps I would use something like:
(defn match [re e]
(re-find re (:a e))
(defn search [re m]
(into #{} (filter (partial match re) m)))
(filter (comp #{"help"} :a) stuff): the freshly-composed function first calls :a on the target, then calls #{"help"} on the result: this returns a truthy value iff the :a attribute is exactly "help".
Converting this into a set, and encapsulating it in a function with the arguments you want to tweak, is left as an exercise for the reader. Frankly, though, the code is so simple that it might well be shorter and more readable to rewrite it each time you want to do a "search".