EuroClojure Learnings: TSV file

tsv_file_load.clj

; My "lightbulb" moment at EuroClojure 2012 was that clojure is at it's most
; powerful when you build abstractions in the data and keep the functions as
; generic as possible so you can chain them easily

; For example, here is some code I wrote a few months ago to parse a freebase TSV file 
(def keywords {"id" "freebase_id"})
(defn override-keywords [field-name]
  	(let [alternative (get keywords field-name)]
     		(if (nil? alternative) field-name alternative)))
(defn get-field-names [line] (map override-keywords (str/split line #"\t")))
(defn get-fields [line] (str/split line #"\t"))
(defn map-fields [hdrs line] (zipmap hdrs line))

(defn map-file [path]
	(with-open [rdr (reader path)]
    (let [lines (into [] (line-seq rdr))
      hdrs (get-field-names (first lines))
      data (map get-fields (rest lines)) ]
    (map #(map-fields hdrs %1) data))))

; On the other hand, here is some code Malcolm Sparks presented that does much the same
; See https://github.com/malcolmsparks/euroclojure2012

(defn get-dataset []
  (let [[header & rows]
        (->> "Downloads/olympic_games.tsv"
             (file (System/getProperty "user.home"))
             reader line-seq
             (map #(split % #"\t")))]
    (map #(zipmap header %) rows)))

; I'm doing a bit more, but there is no doubt that Malcolm's solution is more flexible
; and makes better use of the core library.