xsv 0.3.6 → 0.3.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4c6b42a947e6f518a9284de8262e1c1327267a6e1fdd55ea6d14fffba1f1b19d
-  data.tar.gz: '04887e8c4e50d3bb5762c6738237442d3b6360f4b96304b5a802215d7f4240ca'
+  metadata.gz: f5a4e91e03098400d7c83734fe68128418de719c3c9c0dc95e2e31e5beca9f90
+  data.tar.gz: 78488c6161efcec8e503403b35181b7efe827fb8a67cf8f579d06b444dc50a8f
 SHA512:
-  metadata.gz: e8ad674f6735a1711190469836f30911bb2ebce502d3e99bc268c1c439c7cf9b8f6841189c76950c1722a0861a769a00b6ee7f15e5ab1cb2763a3d6b05342057
-  data.tar.gz: 07c01bc46762348b87f745fd1c7b3aa9d16c0051916b46473a0c0c76bf4d1551f5d2cabc724a612069f3ce83eb959cbe06f4114bef361e615fc37b9ceae51193
+  metadata.gz: e43aa9c8ace8fee462cfaa09a8af785c7522c1ce31810523f5e4d498e3239b8c8e096ce11f638cb31561fc37ed0f7ce9d65d992dbaff23876c1f3fa68d9b44fb
+  data.tar.gz: 812911282e79f7737a27410b71bbe92fc941af181ae7e38c774ced448dcf88643a194e294f9b71862b8cb2eba41b01e0ee755fd4800bcae8ea4a025eefa78a9d

data/CHANGELOG.md

@@ -1,5 +1,14 @@
 # Xsv Changelog
 
+## 0.3.7 - 2020-03-05
+
+Reduce retained memory, making Xsv the definite performance king among the
+Ruby Excel parsing gems.
+
+## 0.3.6 - 2020-03-05
+
+Reduce memory usage
+
 ## 0.3.5 - 2020-03-02
 
 Fix a Gemfile small Gemfile issue that broke the 0.3.3 and 0.3.4 releases

data/lib/xsv/sheet.rb

@@ -28,10 +28,12 @@ module Xsv
     # There is no need to create Sheets from application code.
     #
     # @param workbook [Workbook] The Workbook with shared data such as shared strings and styles
-    # @param io [IO] A handle to an open worksheet XML file or a string with the XML contents
-    def initialize(workbook, io)
+    # @param io [IO] A handle to an open worksheet XML file
+    # @param size [Number] size of the XML file
+    def initialize(workbook, io, size)
       @workbook = workbook
       @io = io
+      @size = size
       @headers = []
       @mode = :array
       @row_skip = 0
@@ -46,11 +48,19 @@ module Xsv
 
     # Iterate over rows, returning either hashes or arrays based on the current mode.
     def each_row(&block)
-      @io.rewind if @io.respond_to?(:rewind)
+      @io.rewind
 
       handler = SheetRowsHandler.new(@mode, empty_row, @workbook, @row_skip, @last_row, &block)
 
-      Ox.sax_parse(handler, @io)
+      # For smaller sheets, memory performance is a lot better if Ox is
+      # handed a string. For larger sheets this leads to awful performance.
+      # This is probably caused by either something in SheetRowsHandler or
+      # the interaction between Zip::InputStream and Ox
+      if @size > 100_000_000
+        Ox.sax_parse(handler, @io)
+      else
+        Ox.sax_parse(handler, @io.read)
+      end
 
       true
     end

data/lib/xsv/sheet_bounds_handler.rb

@@ -17,8 +17,8 @@ module Xsv
         return rows, cols
       end
 
-      sheet.rewind if sheet.respond_to?(:rewind)
-      Ox.sax_parse(handler, sheet)
+      sheet.rewind
+      Ox.sax_parse(handler, sheet.read)
 
       return rows, cols
     end

data/lib/xsv/version.rb

@@ -1,4 +1,4 @@
 # frozen_string_literal: true
 module Xsv
-  VERSION = "0.3.6"
+  VERSION = "0.3.7"
 end

data/lib/xsv/workbook.rb

@@ -80,15 +80,7 @@ module Xsv
       @zip.glob("xl/worksheets/sheet*.xml").sort do |a, b|
         a.name[/\d+/].to_i <=> b.name[/\d+/].to_i
       end.each do |entry|
-        # For smaller sheets, memory performance is a lot better if Ox is
-        # handed a string. For larger sheets this leads to awful performance.
-        # This is probably caused by either something in SheetRowsHandler or
-        # the interaction between Zip::InputStream and Ox
-        if entry.size > 100_000_000
-          @sheets << Xsv::Sheet.new(self, entry.get_input_stream)
-        else
-          @sheets << Xsv::Sheet.new(self, entry.get_input_stream.read)
-        end
+        @sheets << Xsv::Sheet.new(self, entry.get_input_stream, entry.size)
       end
     end
   end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: xsv
 version: !ruby/object:Gem::Version
-  version: 0.3.6
+  version: 0.3.7
 platform: ruby
 authors:
 - Martijn Storck