Merge pull request #244 from CartoDB/equalbins

Convert CDB_EqualIntervalBins to a single SQL statement and add float version

scripts-available/CDB_EqualIntervalBins.sql

@@ -1,8 +1,8 @@
 --
 -- Calculate the equal interval bins for a given column
 --
--- @param in_array A numeric array of numbers to determine the best
---                   to determine the bin boundary
+-- @param in_array An array of numbers to determine the best
+--                   bin boundary
 --
 -- @param breaks The number of bins you want to find.
 --  
@@ -11,27 +11,14 @@
 -- 
 --
 
-CREATE OR REPLACE FUNCTION CDB_EqualIntervalBins ( in_array NUMERIC[], breaks INT ) RETURNS NUMERIC[] as $$
-DECLARE 
-    diff numeric;
-    min_val numeric;
-    max_val numeric;
-    tmp_val numeric;
-    i INT := 1;
-    reply numeric[];
-BEGIN
-    SELECT min(e), max(e) INTO min_val, max_val FROM ( SELECT unnest(in_array) e ) x WHERE e IS NOT NULL;
-    diff = (max_val - min_val) / breaks::numeric;
-    LOOP
-        IF i < breaks THEN
-            tmp_val = min_val + i::numeric * diff;
-            reply = array_append(reply, tmp_val);
-            i := i+1;
-        ELSE
-            reply = array_append(reply, max_val);
-            EXIT;
-        END IF;
-    END LOOP;
-    RETURN reply;
-END;
-$$ language plpgsql IMMUTABLE;
+CREATE OR REPLACE FUNCTION CDB_EqualIntervalBins ( in_array anyarray, breaks INT ) RETURNS anyarray as $$
+WITH stats AS (
+  SELECT min(e), (max(e)-min(e))/breaks AS del
+    FROM (SELECT unnest(in_array) e) AS p)
+SELECT array_agg(bins)
+  FROM (
+    SELECT min + generate_series(1,breaks)*del AS bins
+      FROM stats) q;
+$$ LANGUAGE SQL IMMUTABLE;
+
+DROP FUNCTION IF EXISTS CDB_EqualIntervalBins( numeric[], integer);