Using SAX (Java) to parse multiple XML messages from a single TCP-stream

Question

I'm in a position where I use Java to connect to a TCP port and am streamed XML documents one after another, each delimited with the <?xml start of document tag. An example which demonstrates the format:

<?xml version="1.0"?>
<person>
    <name>Fred Bloggs</name>
</person>
<?xml version="1.0"?>
<person>
    <name>Peter Jones</name>
</person>

I'm using the org.xml.sax.* api. The SAX parsing works perfectly for the first document but throws an exception when it comes across the start of the second document:

Exception in thread "main" org.xml.sax.SAXParseException: The processing instruction 
target matching "[xX][mM][lL]" is not allowed.

The following skeleton class demonstrates the setup I'm using:

import org.xml.sax.InputSource;
import org.xml.sax.XMLReader;
import org.xml.sax.helpers.DefaultHandler;
import org.xml.sax.helpers.XMLReaderFactory;

import java.io.FileReader;

public class XMLTest extends DefaultHandler {

  public XMLTest() {
     super();
  }

  public static void main(String[] args) throws Exception {
    XMLReader xr = XMLReaderFactory.createXMLReader();

    XMLTest handler = new XMLTest();
    xr.setContentHandler(handler);
    xr.setErrorHandler(handler);

    xr.parse(new InputSource(new Socket("127.0.0.1", 4555).getInputStream()));
  }
}

I have no control over the format of the xml (it's a financial data feed), but I need to be able to parse it efficiently, and parse all the documents. I've spent the afternoon/evening trying different things but none have yielded results. Any help would be greatly appreciated.

You have to call parse for each separate document, which means you need to filter and break up the input stream on the ' — Steven D. Majewski, Jul 21 '10 at 18:53
I had to do something like this and just replied (to myself) [here](http://stackoverflow.com/questions/6711766/multiple-xml-files-in-one-stream/) wrapping everything in its own Reader for simpler use — Filipe Pina, Jul 27 '11 at 17:23

score 7 · Accepted Answer · answered Jul 21 '10 at 18:49

7

You'd like to split the stream on every <?xml version="1.0"?> and parse them all separately. The BufferedReader may be helpful in this. Kickoff example:

reader = new BufferedReader(new InputStreamReader(input, "UTF-8"));
StringBuilder builder = null;
for (String line; (line = reader.readLine()) != null;) {
    if (line.startsWith("<?xml")) {
        if (builder != null) {
            xr.parse(new InputSource(builder.toString()));
        }
        builder = new StringBuilder();
    }
    builder.append(line);
}

answered Jul 21 '10 at 18:49

BalusC

1,040,783
362
3,548
3,513

When doing this when `input` is `InputStream input = new Socket("127.0.0.1", 4500).getInputStream();` I get the following exception: Exception in thread "main" java.io.FileNotFoundException: /Users/admin/IdeaProjects/XMLTest/< (No such file or directory) at java.io.FileInputStream.open(Native Method) at java.io.FileInputStream.(FileInputStream.java:106) at java.io.FileInputStream.(FileInputStream.java:66) It seems xr.parse() doesn't like strings, even when wrapped as an InputSource. – jkt Jul 21 '10 at 19:14
Do you consider yourself capable to interpret stacktraces? I don't see how `FileNotFoundException` is related to this all. I'd say, your problem lies somewhere else, maybe in the step beyond parsing. The in the exception message given filename `/Users/admin/IdeaProjects/XMLTest/ – BalusC Jul 21 '10 at 19:18
Hey, I can read stacktraces - I only pasted the first few lines. The stacktrace pointer to my code is `at XMLTest.main(XMLTest.java:42)` and line 42 is: `xr.parse(new InputSource(builder.toString()));` (which is from your example above). I appreciate your assistance with this. – jkt Jul 21 '10 at 19:34
The solution is to wrap the StringBuilder in a StringReader, ie: `xr.parse(new InputSource(new StringReader(builder.toString())));` Thanks for your assistance! – jkt Jul 21 '10 at 19:42

Using SAX (Java) to parse multiple XML messages from a single TCP-stream

1 Answers1

Linked