Merge branch 'master' of github.com:liebke/incanter

modules/incanter-core/src/incanter/core.clj

@@ -2027,6 +2027,58 @@ altering later ones."
                           replace-by-number-or-value old-col-names col-map)]
        (col-names data new-col-names))))
 
+(defn- update
+  ([m key f] (update-in m [key] f))
+  ([m key f & kfs] (apply update (update-in m [key] f) kfs)))
+
+(defn replace-column
+  "Replaces a column in a dataset with new values."
+  ([column-name values]
+     (replace-column column-name values $data))
+  ([column-name values data]
+     (update data :rows
+             (fn [rows]
+               (map #(assoc %1 column-name %2)
+                    rows values)))))
+
+(defn add-column
+  "Adds a column, with given values, to a dataset."
+  ([column-name values]
+     (add-column column-name values $data))
+  ([column-name values data]
+     (if (some #{column-name} (:column-names data))
+       (replace-column column-name values data)
+       (update data :column-names #(conj % column-name)
+               :rows #(mapv (fn [r v]
+                              (assoc r column-name v))
+                            % (concat values (repeat nil)))))))
+
+(defn add-derived-column
+  "This function adds a column to a dataset that is a function of
+  existing columns. If no dataset is provided, $data (bound by the
+  with-data macro) will be used. f should be a function of the
+  from-columns, with arguments in that order.
+
+  Examples:
+    (use '(incanter core datasets))
+    (def cars (get-dataset :cars))
+
+    (add-derived-column :dist-over-speed [:dist :speed] (fn [d s] (/ d s)) cars)
+
+    (with-data (get-dataset :cars)
+      (view (add-derived-column :speed**-1 [:speed] #(/ 1.0 %))))"
+
+  ([column-name from-columns f]
+     (add-derived-column column-name from-columns f $data))
+  ([column-name from-columns f data]
+     (update data :column-names #(conj % column-name)
+             :rows (fn [rows]
+                     (mapv (fn [row]
+                             (assoc row column-name
+                                    (apply f (map #(map-get row %)
+                                                  from-columns))))
+                           rows)))))
+
 ;; credit to M.Brandmeyer
 (defn transform-col
 " Apply function f & args to the specified column of dataset and replace the column

modules/incanter-core/test/incanter/dataset_tests.clj

@@ -51,6 +51,16 @@
   (is (= (head 2 dataset1) ($ (range 2) :all dataset1)))
   (is (= (head dataset1) ($ (range 2) :all dataset1))))
 
+(deftest test-add-column
+  (is (= (add-column :price [17599 22099] cars)
+         (dataset [:speed :weight :colour :price] [(conj car0 17599) (conj car1 22099)])))
+  (is (= (add-column :weight [2500 3500] cars)
+         (dataset [:speed :weight :colour] [(assoc car0 1 2500) (assoc car1 1 3500)]))))
+
+(deftest test-add-derived-column
+  (is (= ($ :weight-minus-speed (add-derived-column :weight-minus-speed [:weight :speed] #(- %1 %2) cars))
+         [5940 6930])))
+
 ;; (deftest selects-on-badly-named-atoms
 ;;   (let [with-nots (dataset [:first :second] [[:not :all] [:all :not]])]
 ;;     (is (= ($ :first