Need suggestion about a fast XML parser

Question

I have a program which needs to parse lots of XML files of sizes varing from few MBs to hundreds of MBs. It needs to parse the file in one-pass, and for which the SAX approach is best suited. I initially tried coding this program using the SAX parser provided by Java library, but it takes way-way too much time to parse the files. I then googled out Piccolo XML Parser, but it throws ArrayIndexOutOfBound exception. Furthermore, sometimes XML files could terminated prematurely.
Can you suggest any SAX XML parser for this job?

Answer 1

I am pasting here a trimmed version of my code..... sorry this is not the simplified and complete in itself version as the norm is.
public Vector<String> search(String dumpfile, String srfor, int type, long size, boolean showTTH, boolean phpSerialize, String hubname) {
     Vector<String> results = new Vector<String>();
     try {
         SAXParserFactory factory = SAXParserFactory.newInstance();
         SAXParser parser = factory.newSAXParser();
         BufferedInputStream ubin = new BufferedInputStream(new FileInputStream(dumpfile));
         ubin.read(new byte[2]); //To discard the starting BZ flag.
         CBZip2InputStream bin = new CBZip2InputStream(ubin);
         String line = null;
         int c=0;
         line = "";
         while((c=bin.read())!='\n' && c!=-1) line = line + (char) c;
         String dnt = line;
         if (c==-1){// (line == null){
          dnt="";
          return null;
         line = "";
         while((c=bin.read())!='\n' && c!=-1) line = line + (char) c;
         if(c==-1)// (line == null)
          return null;
         if (!phpSerialize)
          results.add("Dump's Date and Time stamp: " + dnt + "\n==========================\n"+"hubname: " + line);
         else
          results.add("$" + dnt +"\n"+"|" + line);
         if(hubname!=null && !line.trim().toLowerCase().contains(hubname.toLowerCase().subSequence(0, hubname.length()))){
          results.add("No hits.");
          return results;
         while((c=bin.read())!='\n' && c!=-1);
         FilelistHandler handler = new FilelistHandler(srfor,type,size,showTTH,phpSerialize,hubname,results);
         try{
          parser.parse(bin, handler);
         }catch(org.xml.sax.SAXParseException saxe){
          saxe.printStackTrace();
         if(results.size()==1){
          results.add("No hits.");
     } catch (FileNotFoundException e) {
         // TODO Auto-generated catch block
         e.printStackTrace();
     } catch (IOException e) {
         // TODO Auto-generated catch block
         e.printStackTrace();
     } catch (ParserConfigurationException e) {
         e.printStackTrace();
     } catch (SAXParseException e) {
         // TODO Auto-generated catch block
         e.printStackTrace();
         System.err.println("Line: "+e.getLineNumber()+"; Col:"+e.getColumnNumber());
     } catch (SAXException e) {
         // TODO Auto-generated catch block
         e.printStackTrace();
     return results;
private class FilelistHandler extends DefaultHandler {
     static final int USER = 0;
     static final int FILE = 1;
     static final int DIR = 2;
     static final int ANY = 3;
     static final int UNKNOWN = 4;
     private Vector<String> results;
     Vector<String> dirs = new Vector<String>();
     private String srfor;
     private int type;
     private int level = -1;
     private long size;
     private boolean showTTH;
     private boolean phpSerialize;
     private String hubname;
     private String currentUser;
     private String currentIP;
     public FilelistHandler(String Srfor, int Type, long Size, boolean ShowTTH, boolean PhpSerialize, String Hubname, Vector<String> res) {
         srfor = Srfor.trim().toLowerCase();
         type = Type;
         size = Size;
         showTTH = ShowTTH;
         phpSerialize = PhpSerialize;
         hubname = Hubname;
         results = res;
     public void startElement(String uri, String lname, String qname, Attributes attrs) throws SAXException {
         String result = null;
         long currsize = 0L;
         String TTH = "";
         String value = "";
         int currType = UNKNOWN;
         if (qname.equalsIgnoreCase("Directory")) {
          value = attrs.getValue("Name");
          dirs.add(attrs.getValue("Name"));
          currType = DIR;
         } else if (qname.equalsIgnoreCase("user")) {
          currentUser = attrs.getValue("username");
          currentIP = attrs.getValue("ip");
          currType = USER;
         } else if (qname.equalsIgnoreCase("File")) {
          value = attrs.getValue("Name");
          currsize = Long.parseLong(attrs.getValue("Size"));
          TTH = attrs.getValue("TTH");
          currType = FILE;
         if (currType == FILE || currType == DIR) {
          // Searching.
          boolean found = false;
          result = currentUser + ":" + currentIP + ":" + (currType == FILE && showTTH ? TTH + ":" : "") + getPwd(dirs);
          if (result.trim().toLowerCase().contains(srfor.subSequence(0, srfor.length())) || currType == FILE && TTH.equalsIgnoreCase(srfor)) {
              found = true;
          if (type != ANY && currType != type)
              found = false;
          if (size >= 0 && currType == FILE)
              if ((size == 0 && currsize != size) || (size != 0 && (((double) (Math.abs(currsize - size))) / size) > 0.1))
               found = false;
          if (found) {
              if (!phpSerialize) {
               if (currType == FILE)
                   result = result + "/" + value;
              } else {
               int index = 0;
               result = serializeEntity(index++, currType == FILE ? "f" : "d");
               result = result + serializeEntity(index++, currentUser);
               result = result + serializeEntity(index++, currentIP);
               if (currType == FILE && showTTH)
                   result = result + serializeEntity(index++, TTH);
               result = result + serializeEntity(index++, getPwd(dirs));
               if (currType == FILE)
                   result = result + serializeEntity(index++, value);
               result = "a:" + index + ":{" + result + "}";
              results.add(result); // ADDING THE RESULT.
     public void endElement(String uri, String lname, String qname) throws SAXException {
         if (qname.equalsIgnoreCase("Directory")) {
          dirs.remove(dirs.size() - 1);
     private String getPwd(Vector<String> dirs) {
         String pwd = "";
         for (String dir : dirs) {
          pwd = pwd + "/" + dir;
         return pwd;
     private String serializeEntity(int index, String s) {
         return "i:" + index + ";s:" + s.length() + ":\"" + s + "\";";
     public Vector<String> getParsedData() {
         return results;
    }