"David Jessee" <DavidJessee@discussions.microsoft.com> wrote in message
news:D1ABF838-AF0E-4B7A-AC75-958DFF78FC2B@microsoft.com...
>I have a schema that I've used to generate a class structure vis the
>xsd.exe
> utility. I'm having to manually validate the resultant XML since the
> resultant document is not necessarily compliant with the structure defined
> by
> the original XSD.
>
> Here's what I'm wondering, though.....
> I have a server process that is going to have to process these documents
> rather quickly. At the end of the document, there are going to be
> attached
> files (base64encoded). During this processing, there isn't any sense in
> loading up the entire XML statement into a DOM document because I only ned
> the first few nodes in order to determine what to do.
>
> The logical choice: Use a SAX parser (XmlReader).
>
> My original XSD defined a number of complex types that consisted of
> sequences of elements. According to a strict interpretation of the schema
> (which would help SAX parsing) those elements would have to occur in that
> exact order. This would REALLY simplify parsing the document via SAX. I
> don't see anything in the generateds class that indicates an element's
> cardinality, though.
>
> Can I assume that even though the original schema stated that a sequence
> of
> elements must ccur in a specified order, that does not necessarily mean
> that
> the XmlSerializer will honor that definition?
> Is there a way to control this, or do I have to take the route of loading
> the entire document into a XmlDocument (in memory) instead of just pulling
> the information I need out of a stream?