Clojure string replace on a vector of maps with text

Question

I want to replace some characters in a vector of maps with a text in it.

This should be a part of a bigger program wich counts all the words in a list of texts.

The input-vector looks like this:

[{:text "bla. Bla! Blabla, foo"}
   {:text "hello foo? bla Foo, blabla"}
   {:text "bla blub Foo Bla blub"}]

The output should looks like this and should be sorted on the values:

{:bla 3 :Bla 2 :blub 2 :foo 2 :Foo 2 ... }

But first I want too clean the strings from some characters.

I tried it with map but I don't understand why this code is not working right:

(defn clean-texts []
  (map (fn [x] (clojure.string/replace x #"[.,]" "")) (:text texts)))

The whole code looks like this:

(ns keyword-finder.core
  (:gen-class))

(def texts
  [{:text "bla. Bla! Blabla, foo"}
   {:text "hello foo? bla Foo, blabla"}
   {:text "bla blub Foo Bla blub"}])

(defn clean-texts []
  (map (fn [x] (clojure.string/replace x #"[.,]" "")) (:text texts))
)

When you pose a question about code that does not work, it helps immensely if you take the time to describe what you expect to happen, and what's happening instead. — Pointy
– Pointy, Commented Oct 12, 2014 at 13:59
What results do you get from running clean-texts and why are they incorrect? — soulcheck
– soulcheck, Commented Oct 12, 2014 at 14:20

Diego Basch · Accepted Answer · 2014-10-12 15:35:18Z

5

What you want is something like this:

(defn tokenize [s]
  (-> s
    (.replaceAll "[^a-zA-Z\\s]" "")
    (clojure.string/split #" ")))

This removes all non-letters from a string, so when applied to "bla. blah, blah" it will give you "bla blah blah"

(defn word-counts [texts]
  (let [tokens
    (->> texts
        (map (comp tokenize :text))
        (apply concat)
        (map keyword))]
   (frequencies tokens)))

This function extracts the values for the key :text from your map, applies tokenize to all resulting strings, concatenates them into a list of words, converts them into keywords, and finally returns the keyword counts using the built-in function frequencies.

(word-counts texts)

produces {:bla 3, :Bla 2, :Blabla 1, :foo 2, :hello 1, :Foo 2, :blabla 1, :blub 2}

edited Oct 12, 2014 at 15:35

answered Oct 12, 2014 at 15:02

Diego Basch

13.1k2 gold badges31 silver badges24 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Kingalione Over a year ago

Yes but the regex should be "[a-zA-z]\\s". But still thank you

Kingalione Over a year ago

And to make it perfect how can I sort it by values but still returning the same result? Do I have to write a own comparator?

soulcheck · Accepted Answer · 2014-10-12 14:44:40Z

4

You're applying map to the wrong sequence:

(:text texts)

returns nil since :text is applied to the whole texts list.

What you probably wanted to do was to map the inner function on the whole texts list, while extracting :text for each element:

(defn clean-texts []
     (map (fn [x] (clojure.string/replace (:text x) #"[.,]" "")) texts))

answered Oct 12, 2014 at 14:44

soulcheck

36.9k6 gold badges95 silver badges91 bronze badges

Collectives™ on Stack Overflow

Clojure string replace on a vector of maps with text

2 Answers 2

2 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related