Here, the example scripts pull a list of all the nodes with the name "MEDIA".
Using Regular Expressions, you can identify each "MEDIA" node regardless of its letter case.

 
var doc:XML=
<DATA>
	<CATEGORY>
		<MEDIA>A.jpg</MEDIA>
		<MEDIA>B.jpg</MEDIA>
	</CATEGORY>
	<CATEGORY>
		<MEDIA>C.jpg</MEDIA>
	</CATEGORY>
	<CATEGORY>
		<MEDIA>D.jpg</MEDIA>
	</CATEGORY>
	<CATEGORY>
		<MEDIA>E.jpg</MEDIA>
	</CATEGORY>
</DATA>;
 
var findName:String="media";
trace(doc..*.(new RegExp(findName,"i")["test"](name())).toXMLString());
 

"Error #1069: Property lowerCase not found on QName"

 
trace(doc..*.(name().lowerCase()=="media").toXMLString());
 

Letter case must be same for this to work.

 
trace(doc..*.(name()=="MEDIA").toXMLString());
 

Used a regular expression to do case-insensitive filter
Works, letter case does not matter...but the expression is too "hardcoded".

 
trace(doc..*.(/media/i.test(name())).toXMLString());
 

Works, but compiler does not like it:
"Warning: 3594: test is not a recognized method of the dynamic class RegExp."
This is an attempt to allow variables and expressions.

 
var findName:String="media";
trace(doc..*.(new RegExp(findName,"i").test(name())).toXMLString());
 

:jumping: Works!
"[" and "]" keeps the compiler from warning when using the RegExp "test()" method.
Now letter case does not matter, and RegExp constructor allows variables or expressions.

 
var findName:String="media";
trace(doc..*.(new RegExp(findName,"i")["test"](name())).toXMLString());
 

 
<library>
	<book>
<page id="one"/>
		<worm name="Eddy" />
<page id="two"/>
<page id="three">
			<worm name="Lisa"/>
			<worm name="Pete" />
		</page>
<page id="four"/>
	</book>
</library>
 

There are some E4X operators and XML methods to help use do this when combined together:

.. Access all descendants of the XML object.
. Access a property of an Object or XML.
* Wildcard, match any part of the XML.
childIndex() Index number of an XML object relative to its parent node.
name() Name of the XML node.

Return a certain "worm" node at the index it belongs to from its XMLList.
Any nested "worm" node occurrence is also considered.

 
var doc:XML=
<library>
	<book>
<page id="one"/>
		<worm name="Eddy" />
<page id="two"/>
<page id="three">
			<worm name="Lisa"/>
			<worm name="Pete" />
		</page>
<page id="four"/>
	</book>
</library>;
 
//Pick the "worm" node at index 1.
var data:*=doc..*.worm[1];
trace(data.toXMLString());
 
/*=========================================
OUTPUT:
 
<worm name="Lisa"/>			//worm[1]
===========================================*/
 

Delete all "worm" nodes from the xml, even the nested ones.

 
var doc:XML=
<library>
	<book>
<page id="one"/>
		<worm name="Eddy" />
<page id="two"/>
<page id="three">
			<worm name="Lisa"/>
			<worm name="Pete" />
		</page>
<page id="four"/>
	</book>
</library>;
 
delete doc..*.worm;
trace(doc.toXMLString());
 
/*=========================================
OUTPUT:
All "worm" nodes are deleted.
<library>
  <book>
<page id="one"/>
<page id="two"/>
<page id="three"/>
<page id="four"/>
  </book>
</library>
===========================================*/
 

Use E4X filtering with Regular Expressions.
XML methods and Regular Expressions can be used inside parentheses to accomplish E4X filtering.

doc..*.(childIndex()==0 && /library|book/i.test(name())==false)

List all nodes that belong at a given index of their XMLList.

 
var doc:XML=
<library>
	<book>
<page id="one"/>
		<worm name="Eddy" />
<page id="two"/>
<page id="three">
			<worm name="Lisa"/>
			<worm name="Pete" />
		</page>
<page id="four"/>
	</book>
</library>;
 
//----------------------------------------------------
//Get the first node of each list whose name does not
//match the regular expression.
//----------------------------------------------------
var data:*=doc..*.(childIndex()==0 && /library|book/i.test(name())==false);
trace(data.toXMLString());
 
/*=========================================
OUTPUT:
<page id="one"/>
<worm name="Lisa"/>
===========================================*/