HTML Stripping -- too much memory

Question

I have a program that takes data from a database and strips it of html tags. This is done over multiple tables. Some of the tables return a few thousand lines, but then others return about 65,000 lines. I open the file created and clean the needed fields with the following method. This causes an Out of Memory error with some of the larger files. Is there a better way to strip the html tags than I am using that would not throw this Out of Memory error?
      * Removes the html formatting for a string and stores it in a file
      * @return java.lang.String
      * @param htmlText java.lang.String
      * @param outputText java.lang.String
     public static String removeHtml(String htmlText) throws IOException     
          String returnValue = "";
          EditorKit kit = new HTMLEditorKit();          
          Document doc = kit.createDefaultDocument();          
          // The Document class does not yet handle charset's properly.          
          doc.putProperty("IgnoreCharsetDirective", Boolean.TRUE);          
          try          
               // Create a reader on the HTML content.     
               Reader rd = new StringReader(htmlText);
               // Parse the HTML.               
               kit.read(rd, doc, 0);               
               //Set returnValue to cleaned text
               returnValue = doc.getText(0, doc.getLength()).toString();
               rd.close();
               doc = null;               
               kit = null;
          catch (Exception e)          
               System.out.println ("ERROR: " + e.getMessage());
               e.printStackTrace();          
          return returnValue;
     }

Answer 1

There is a fairly trivial way where you just copy the characters from the input to the output, except for the following rules:
1. When the input character is '<', stop writing characters to the output (including this one).
2. When the input character is '>', start writing characters to the output (after this one).
However this may stumble on malformed HTML, for example if the '<' character appears unescaped in text. I don't know if your data suffers from this problem or not.