Error Parsing XML using XMLTextReader

[Hang Cheng]

I've been trying to parse the following XML file with the .NET XMLTextReader:

<?xml version="1.0"?><level1><level2><level3>Test</level3><level3>Test2</level3></level2><level2><level3>Test</level3><level3>Test2</level3><level3><level4>TEST<level5>TEST5</level5>TESTP</level4><level4>Test4</level4></level3></level2></level1>

However, the resulting output does not seem to make any sense:
Start Element: level1 @ 0
Start Element: level2 @ 1
Start Element: level3 @ 2
CDATA: Test @ 2
Start Element: level3 @ 2
CDATA: Test2 @ 2
CDATA: @ 1
Start Element: level2 @ 1
Start Element: level3 @ 2
CDATA: Test @ 2
Start Element: level3 @ 2
CDATA: Test2 @ 2
Start Element: level3 @ 2
Start Element: level4 @ 3
CDATA: TEST @ 4
CDATA: TEST5 @ 4
CDATA: TESTP @ 3
Start Element: level4 @ 3
CDATA: Test4 @ 3
CDATA: @ 2
CDATA: @ 1
CDATA: @ 0

Notice how "TEST" and "TESTP" are both contents of the "<level4>" tag, yet XMLTextReader thinks that they are at different levels. Furthermore, the "<level5>" tag is completely ignored as a "Start Element". What's going on? I thought mixing contents in an element was allowed.

Relevant Portion of my code:
public void Parse2()
{
while(MyXMLReader.Read())
{
if (MyXMLReader.IsStartElement())
{
Console.WriteLine("Start Element: " + MyXMLReader.Name + " @ " + MyXMLReader.Depth);
}
else
{
Console.WriteLine("CDATA: " + MyXMLReader.ReadString() + " @ " + MyXMLReader.Depth);
}
}
}

Thanks.

[Oleg Tkachenko [MVP]]

Hang Cheng wrote:

Notice how "TEST" and "TESTP" are both contents of the "<level4>"
tag, yet XMLTextReader thinks that they are at different levels.
Furthermore, the "<level5>" tag is completely ignored as a "Start
Element". What's going on? I thought mixing contents in an element
was allowed.

Well, that's how you read it. To get better understanding of how
XmlReader reads the document, try something simpler:

while(MyXMLReader.Read()) {
Console.WriteLine("Node: {0}, name: {1}, value: {2}, depth: {3}",
MyXMLReader.NodeType, MyXMLReader.Name, MyXMLReader.Value,
MyXMLReader.Depth);
}

Your particular mistake is that you ignore еур behaviour of ReadString()
method when reading mixed content. As stated in MSDN:

"If positioned on an element, ReadString concatenates all text,
significant white space, white space, and CDATA section nodes together
and returns the concatenated data as the element content. It stops when
any markup is encountered. This could occur in a mixed content model, or
when an element end tag is read."

So, if you run the code in debugger, you can see that after calling
ReadString() on level4 node, reader gets positioned on level5 start tag.
But then Read() screws things moving the reader to "TEST5" text node.

--
Oleg Tkachenko [XML MVP, XmlInsider]
http://blog.tkachenko.com