DOM String output is not the same as original XML

[Jim Janssens]

I'm trying to extract an element (and all its children) from a given XML file as string. However, I want to retain its current infoset. Bassically, I want the parser to leave alone the XML I'm extracting.

So, to do this I build a DOM document from a given XML string, I search for the desired element, and next I output that element via a Transformer to a String.

The problem is that the orignal XML is not the same any more as the outputed XML

The source XML might be like this:

<envelope xmlns="http://envelope.com>
<person xmlns="http://person.com">
    <ns1:name xmlns="http://person.com" 
   xmlns:ns1="http://identity.com/>
    <ns1:firstName xmlns="http://person.com"
   xmlns:ns1="http://identity.com/>
</person>
</envelope>

I extract the 'person' element and write it back to String, then it looks like this:

<person xmlns="http://person.com">
    <ns1:name  xmlns:ns1="http://identity.com/>
    <ns1:firstName xmlns:ns1="http://identity.com/>
</person>

Now, I now that the input XML contains redundant namespacing, and that the output XML is better. But in this case I want to output XML exactly to be as the input XML:

- Whitespaces/tabs/linefeeds/cariaged/what ever must be retained
- No namespace optimalisation whatsoever

So, what I'm looking for a 'substring()' on the original XML. The problem is that real substring is not that simple on an XML and probably the least prefered/clean solution.

I tried the 'Transform' class from xalan, you can configure a lot there. I managed to configure it so it leaves the indenting alone. But it still does namespace optimalisation and also removes linefeeds between namespace declarations. If I would have

<someElement xmlns:ns1="test1"
xmlns:ns2="test2"</someElement>

Then the output looks like:

<someElement xmlns:ns1="test1" xmlns:ns2="test2"</someElement>


Any advise is welcome !

[William Brogden]

Parsers and Transformers are concerned with creating legal DOMs and writing legal XML documents, not with conserving exact formats. For example there is no guarantee that the order of attributes will be preserved and of course there is the well known example that there are two legal ways to write an empty element.

Thats just a fact of life. If you want a particular exact format you are probably going to have to create it yourself.

Bill

[Henry Wong]

So, what I'm looking for a 'substring()' on the original XML. The problem is that real substring is not that simple on an XML and probably the least prefered/clean solution.

If all you want is to extract substrings, and the fields that you are trying to extract are simple (meaning a field doesn't nest another field with the same element name), you can use regular expressions to extract the substring. Try...

public static String[] fetchAllSimpleXmlByTag(String origxml, String tag) {
  Pattern p = Pattern.compile("(?s)(<(" + tag + ")\\b.*?>.*?</\\2>)");
  Matcher m = p.matcher(origxml);
 
  ArrayList<String> results = new ArrayList<String>();
  while (m.find()) {
    results.add(m.group(1));
  }
  return results.toArray(new String[0]);
}

Henry