How do I skeletonize XML using the DOM object?

[Mitch_H]

I have thousands of variations of the ubber-simplified XML listed below which I must "Skeletonize" and create an output version of the new transformed xml. (Spaces replaced with dots to maintain indentation for readability)

</sampleDoc>
  <stateCounty>
    <state abbreviation="CA" fips="06">
      <name>California</name>
    </state>
    <county code="RV" fips="065">
      <name>Riverside</name>
    </county>
  </stateCounty>
  <buildInfo>
    <date>20120529</date>
    <time>11.08.00</time>
  </buildInfo>
  <userInfo>
    <userId>
      <id securityLevel="23">12345678</id>
      <name>Fred Flintstone</name>
    </userId>
  </userInfo>
</sampleDoc>

I need to remove some of the elements, blank out others, and blank out some attributes, to create the resulting skeleton XML which still has the <stateCounty> element and all it's children, deleted the <buildInfo> element and children, and removed the values from the attributes and element data in <userInfo>'s children as shown below.

</sampleDoc>
  <stateCounty>
      <state abbreviation="CA" fips="06">
      <name>California</name>
    </state>
    <county code="RV" fips="065">
      <name>Riverside</name>
    </county>
  </stateCounty>
  <userInfo>
    <userId>
      <id securityLevel=""></id>
      <name></name>
    </userId>
  </userInfo>
</sampleDoc>

I am traditionally a mainframe programmer and am fairly new to Java. I have been searching for information on how to use the DOM object, but don't understand how to determine the difference between the /sampleDoc/stateCounty/state/name and /sampleDoc/userInfo/userId/name elements? They both seem to have the same element name "name".

If anyone can provide any insight on how to process this tree, your assistance would be greatly appreciated.

Mitch Harper

[Tolls]

If you have anyone with xslt (ie xml transformation) knowledge then I would simply transform the former into the latter.
It's a pretty simple one (assuming that is the actual data format).

If not, and you need to do this via Java, then you might want to use a SAX parser, which will let you trace down the tree.
Failing that (and since you seem to be looking at DOM) you could use XPath.

[JosAH]

Because I'm a lazy old sod I'd use a SAXParser and use an extension of a DefaultHandler object; that DefaultHandler is used as a callback function when the SAXParser parses the XML text. No fat DOM model is created and you can create your skeleton output while the input is being parsed; all lean and mean ;-)

kind regards,

Jos

[Tolls]

I'd transform it myself.
But then I did shed loads of that a few years back, so it's almost second nature.

The DOM is a problem, especially if the source is big.

[Mitch_H]

Tolls and JosAH,

Thanks for your advise. I was able to tackle the beast using the SAX Parser as recommended.

Your advise is much appreciated.

Mitch Harper