Mimic BufferedReader with XML File

Question

I have a class (extends BufferedReader) that reads "records" 1 by 1 -
import java.io.*;
import java.util.*;
public class FileRecordReader extends BufferedReader {
  public FileRecordReader(Reader reader) throws FileNotFoundException {
    super(reader);
  public String[] readRecord() throws IOException {
    Vector fields = new Vector();
    char[] buffer  = new char[1];
    int bytesRead;
    StringBuffer newField = new StringBuffer();
    bytesRead = read(buffer);
    while (bytesRead != -1) {
      // ignore linefeed and carridge returns
      if (buffer[0] == '\n' || buffer[0] == '\r') {
        bytesRead = read(buffer);
     continue;
      if (buffer[0] == '|') {
        fields.addElement(newField.toString());
     newField = new StringBuffer();
     bytesRead = read(buffer);
     continue;
      if (buffer[0] == '~') {
        break;
      newField.append(buffer[0]);
      bytesRead = read(buffer);
    if (fields.size() == 0) return null;
    String[] returnFields = new String[fields.size()];
    fields.copyInto(returnFields);
    return returnFields;
}As you can see the fields are seperated by the '|' character and the '~' character signals the end of a record.
I now have to extend this class to handle xml files as well. In the readRecord method I check a property which tells me what file type I'm dealing with (txt file or xml) and then I call either readDelimitedRecord or readXMLRecord.
readDelimitedRecord does exactly what readRecord currently does, and readXMLRecord should do the equivalent for an xml file. I know how to read a record from the xml file, but how do I keep track of what records I've read so far?

Answer 1

Your solution for delimited files doesn't scale to XML files. Besides, XML files don't have "records". I would recommend that for XML files you do something completely different. Feed them to a SAX parser and get the stream of startDocument(), startElement(), characters(), endElement(), and endDocument() events. Otherwise you would find yourself writing a whole XML parser, which is more complicated than it looks.