add test case for complex rowspan+colspan

siuying · Jan 26, 2010 · 608686d · 608686d
1 parent 5437ced
commit 608686d
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 2 deletions.
diff --git a/lib/table_parser.rb b/lib/table_parser.rb
@@ -4,5 +4,5 @@
 require 'table_parser/parser'
 
 module TableParser
-  VERSION = '0.5.6'
+  VERSION = '0.5.7'
 end
diff --git a/lib/table_parser/parser.rb b/lib/table_parser/parser.rb
@@ -65,7 +65,11 @@ def self.extract_nodes(rows, headers, dup_rows, dup_cols)
               end
             end
 
-            if col.rowspan > 1 && data[row_index+1]
+            if col.rowspan > 1
+              unless data[row_index+1]
+                data.insert(row_index, [])
+              end
+
               if dup_rows
                 data[row_index+1].insert(col_index, TableNode.new(col.element, col.rowspan - 1))
               else

diff --git a/test/test_table_parser.rb b/test/test_table_parser.rb
@@ -101,11 +101,31 @@ def test_parse_noheader
     assert_equal(3, table[1].size)
   end
 
+  def test_parse_complex_colrowspan
+    html = "<html><body><table><tr><td>A</td><td>B</td><td>C</td><td>D</td><td>E</td></tr>\
+      <tr><td rowspan=\"5\">1</td><td>2</td><td>3</td><td>4</td><td>5</td></tr> \
+      <tr><td rowspan=\"2\">2b</td><td>3d</td><td>4d</td><td>5b</td></tr> \
+      <tr><td>3c</td><td rowspan=\"2\">4c</td><td>5c</td></tr>\
+      <tr><td rowspan=\"2\">2d</td><td>3d</td><td>5d</td></tr>\
+      <tr><td>3e</td><td>4e</td><td>5e</td></tr>\
+      </table></body></html>"
+
+    doc = Nokogiri::HTML(html)
+    table = TableParser::Table.new doc, "/html/body/table", {:dup_cols => false, :dup_rows => false}
+    puts table
+    assert_equal(5, table.columns.size, 'header_count should = 5 ')
+    assert_equal(1, table[0].size)
+    assert_equal(3, table[1].size)
+    assert_equal(5, table[2].size)
+    assert_equal(4, table[3].size)
+    assert_equal(5, table[4].size)
+  end
 
   def test_web
     html = open("test4.html").read
     doc = Nokogiri::HTML::Document.parse(html, nil, "Shift_JIS")
     table = TableParser::Table.new doc, "/html/body/div/div[3]/div/div[2]/table", {:header => false, :dup_rows => false}
     puts table.columns[0].size
   end
+
 end