SAX XMLReader, XMLFilter, ContentHandler and XMLWriter question

[Jeff Calico]

Hello all. I am implementing a SAX filter to strip a bunch of unneeded
elements out of a large XML file. I found a book "Java & XML" by Brett
McLaughlin, and an interesting article by him wich address my issues:
http://www-128.ibm.com/developerwork...ipbigdoc3.html

However, doing it in that specified way does not seem to work at all!
Doing it very differently *seems* to work, but most likely I am not
understanding something.

Here is my code, contrasted with the book's code. The class
KeepSpecificEle mentsFilter is at the end:

--------------------------
MY TRIAL AND ERROR WAY:
---------------------------
FileReader r = new FileReader( "filename" );
XMLReader xr = XMLReaderFactor y.createXMLRead er();
KeepSpecificElt sFilter filter = new KeepSpecificElt sFilter( xr,
"elt");
XMLWriter xw = new XMLWriter( filter, new FileWriter( "Out.xml" ) );
xw.parse( new InputSource(r) );

--------------------------------------------------------
THE WAY THE BOOK SAYS TO DO IT (maybe I misunderstand):
---------------------------------------------------------
FileReader r = new FileReader( s );
XMLReader xr = XMLReaderFactor y.createXMLRead er();
XMLWriter xw = new XMLWriter( xr, new FileWriter( "jeffOut.xm l" ) );
KeepSpecificElt sFilter filter = new KeepSpecificElt sFilter( xw,
"elt");

//DefaultHandler dh = new DefaultHandler( );
JeffContentHand ler dh = new JeffContentHand ler(xr);

filter.setConte ntHandler( dh );
filter.parse( new InputSource(r) );
------------------------------------------------------------
Note the difference between who does the parsing
(writer or filter) and the way they are chained together.

And Last, here is the Filter class:
------------------------------------------------------------

public class KeepSpecificElt sFilter extends XMLFilterImpl {

private List elementsToKeep;

private boolean inKeptElement = false;

public KeepSpecificElt sFilter( XMLReader parent, String
elementToKeep )
{
super( parent );
elementsToKeep = new LinkedList();
elementsToKeep. add(elementToKe ep);
}

//---------------------------------------------------------------------------

public KeepSpecificElt sFilter( XMLReader parent, List elementsToKeep
)
{
super(parent);
this.elementsTo Keep = elementsToKeep;
}

//---------------------------------------------------------------------------

public void startElement( String uri, String localName, String qName,
Attributes atts)
throws SAXException
{
if( elementsToKeep. contains(localN ame) )
{
System.out.prin tln("In kept element = " + localName);
super.startElem ent( uri, localName, qName, atts );
inKeptElement = true;
}
else
{
}
}

//---------------------------------------------------------------------------

public void endElement( String uri, String localName, String qName )
throws SAXException
{
if( elementsToKeep. contains(localN ame) )
{
super.endElemen t( uri, localName, qName );
inKeptElement = false;
}
else
{
// DON'T DO ANYTHING... PREVENTS PROCESSING OF
ELEMENTS
}
}

//---------------------------------------------------------------------------

public void characters( char ch[], int start, int len )
throws
SAXException
{
if( inKeptElement )
{
super.character s( ch, start, len );
}
}
}

Any insight would be appreciated!

--Jeff

[Jeff Calico]

I forgot to add that I don't understand what to do with the
ContentHandler class;
I tried to use the DefaultHandler, and then I tried my own class
"JeffContentHan dler"
with an empty implementation. It seems to me that the Filter class is
doing this
work though, so why would I register a ContentHandler?

--Jeff

[Joseph Kesselman]

Jeff Calico wrote:

I forgot to add that I don't understand what to do with the
ContentHandler class;
I tried to use the DefaultHandler, and then I tried my own class
"JeffContentHan dler"
with an empty implementation. It seems to me that the Filter class is
doing this
work though, so why would I register a ContentHandler?

Normally, the filter is a ContentHandler whose only job is to pass
selected events along to another ContentHandler which actually uses the
filtered document. You have to register your "real" ContentHandler with
the filter so it knows what to do with the events after deciding whether
to keep them or not.

Alternatively, of course, you can combine both the filtering and the
operate-on-the-data stages in a single custom ContentHandler. But in
that case there's no need for it to claim to be a Filter.
--
Joe Kesselman / Beware the fury of a patient man. -- John Dryden