Here's a quick version in Clojure. I'm sure it can be done cleaner than this..

  (def articles
    [{:title "title1" :author "author1" :tags #{:Java :t2 :t3}}
     {:title "title2" :author "author1" :tags #{:Jvxa :t2 :t3}}
     {:title "title3" :author "author3" :tags #{:Java :t3}}])
  
  ;; find the first article in the collection that has the tag “Java”.
  (first (filter #(contains? (:tags %) :Java) articles))
  ;; ==> {:tags #{:t2 :Java :t3}, :title "title1", :author "author1"}
  
  ;;get all the elements that match instead of just the first
  (filter #(contains? (:tags %) :Java) articles)
  ;; ==> ({:tags #{:t2 :Java :t3}, :title "title1", :author "author1"}
          {:tags #{:Java :t3}, :title "title3", :author "author3"})
  
  ;;group all the articles based on the author.
  (group-by :author articles)  ;; cheating?
  ;; ==> {"author1"
             [{:tags #{:t2 :Java :t3}, :title "title1", :author "author1"}
              {:tags #{:Jvxa :t2 :t3}, :title "title2", :author "author1"}],
          "author3"
             [{:tags #{:Java :t3}, :title "title3", :author "author3"}]}
  
  ;;find all the different tags used in the collections
  (apply clojure.set/union (map :tags articles))
  ;; ==> #{:Jvxa :t2 :Java :t3}