arc-furnace 0.1.30 → 0.1.31

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: d2faf523ade797107093de338836f4cf729280b6
-  data.tar.gz: d44a2695905ba74e5ff6dde9900c093fec51698a
+  metadata.gz: c35f10203b2279c62249355a6c9fe3b5bea5926f
+  data.tar.gz: 7b028a8348e8a07193488436fdb4b7f0375172a5
 SHA512:
-  metadata.gz: 56f4db95b3f40082f98ceaa065f42267b2ba09cb62eba19a18df1a908f1fd95a7ee36ead23e9be508ed1e09f85ac501cf7cb1707212cfa279908ae54d9032975
-  data.tar.gz: aaa3aaf2b622b2846b887c3f1cd67a014d14c48bc7ce27d7cd354c9b530307c8eb72c8db9e49fa4efda8950173c6ec82b1b187194597991c1623dc2d1e6c9d1c
+  metadata.gz: bbe2cc9797ca5544d0d75cb873a912b2150323b7f8e6d4ed67940d859eca9a0a75643f9231fb63c16d8579c466835d6d18a1167f649e78c77bfa8a90bb7c333b
+  data.tar.gz: 567a87f7cbc709875a9522d98dc859529af0cc82d8cfa590f06deba7553c86b0d8439eff7f4aa9ca678b41ceecd2a69c90448dd9f8603d0d5cfe0bd8d349bf1c

data/lib/arc-furnace/excel_source.rb

@@ -5,22 +5,35 @@ module ArcFurnace
   class ExcelSource < EnumeratorSource
 
     private_attr_reader :excel, :header_row
-    attr_reader :value
+    attr_reader :value, :group_by, :key_column
 
-    def initialize(filename: , sheet: nil)
+    def initialize(filename:, sheet: nil, group_by: false, key_column: nil)
       @excel = Roo::Excelx.new(filename)
-      if sheet
-        excel.default_sheet = sheet
-      end
+      @preprocessed_excel = []
+      @group_by = group_by
+      @key_column = key_column
+      excel.default_sheet = sheet if sheet
       super()
     end
 
-    def close
-      @excel.close if @excel
-    end
+    alias_method :group_by?, :group_by
 
     def preprocess
-      enumerator.next
+      if group_by?
+        build_headers
+        group_rows
+      else
+        enumerator.next
+      end
+    end
+
+    def build_headers
+      @header_row = excel.first
+    end
+
+    def group_rows
+      @excel.each_row_streaming { |row| @preprocessed_excel << build_row(row) }
+      @preprocessed_excel = @preprocessed_excel.group_by { |row| row[key_column] }
     end
 
     def extract_cell_value(cell)
@@ -30,21 +43,36 @@ module ArcFurnace
       end
     end
 
+    def build_row(row)
+      row.each_with_object({}) do |cell, result|
+        value = extract_cell_value(cell)
+        result[header_row[cell.coordinate.column - 1]] = value if value
+      end
+    end
+
     def build_enumerator
       Enumerator.new do |yielder|
-        excel.each_row_streaming do |row|
-          yielder <<
-              if header_row
-                row.each_with_object({}) do |cell, result|
-                  value = extract_cell_value(cell)
-                  result[header_row[cell.coordinate.column - 1]] = value if value
-                end
-              else
-                # First time, return the header row so we can save it.
-                @header_row = row.map { |value| extract_cell_value(value) }
+        if group_by?
+          @preprocessed_excel.each { |_, array| yielder << array }
+        else
+          excel.each_row_streaming do |row|
+            yielder << if header_row
+              row.each_with_object({}) do |cell, result|
+                value = extract_cell_value(cell)
+                result[header_row[cell.coordinate.column - 1]] = value if value
               end
+            else
+              # First time, return the header row so we can save it.
+              @header_row = row.map { |value| extract_cell_value(value) }
+            end
+          end
         end
       end
     end
+
+    def close
+      @excel.close if @excel
+    end
+
   end
 end

data/lib/arc-furnace/version.rb

@@ -1,3 +1,3 @@
 module ArcFurnace
-  VERSION = "0.1.30"
+  VERSION = "0.1.31"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: arc-furnace
 version: !ruby/object:Gem::Version
-  version: 0.1.30
+  version: 0.1.31
 platform: ruby
 authors:
 - Daniel Spangenberger
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2018-03-29 00:00:00.000000000 Z
+date: 2018-04-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: msgpack
@@ -230,7 +230,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.6.13
+rubygems_version: 2.4.5.1
 signing_key: 
 specification_version: 4
 summary: Melds and transforms data from multiple sources into a single stream