Newbie: parsing simple XML with C/C++

[Steven Feil]

I'm looking for simple examples of XML parsing using C/C++ that could be applicable to both Unix and Windows programming. I am wanting to parse an XML structure that is basically flat. The Information that I wish to extract would be held within XML attributes. It is important that the order of the document be preserved in the parsing process.

Here is a document similar to the type of XML that will be generated.

<collector name="John Tomas">
<cd title="Moon Light City" artist="Fred Ziffle" />
<book title="Hiking Big Bend" author="Laurence Parent" pages="171" /> <cd title="Ups and Downs" artist="Hank Kimble" /> <cd title="Eat My Shorts" artist="Bart Simpson" /> <book title="100 things to cook" author="Tom Duly" pages="100" /> </collector>

My program would need to know what order the information was in. Specifically the cd "Moon Light City" was found first followed by the book "Hiking Big Bend" and so forth. I would also need to extract all the data, noting that the data for a cd is different than that for a book.

I've looked into xpath, but what I've read about it so far seams much more complicated than what I need. Most of the XML books I've seen seem to concentrate on the DTD and interoperability between HTML and XML. I've written a DTD that I believe is right, but I haven't tested it yet.

Any websites or steering on what I should search the Internet for would be greatly appreciated.

 

[strajan]

You can download xerces parser from apache, which comes with samples on XMLparsing. http://xml.apache.org/xerces-c/index.html

HTH

 

[Patrick TJ McPhee]

% I'm looking for simple examples of XML parsing using C/C++ that could % be applicable to both Unix and Windows programming.

You should use an XML parser, rather than trying to write code that does the parsing. There are a bunch of them out there -- my current favourite is libxml (http://xmlsoft.org), but my old favourite (expat -- don't have a URL) works well and is a little lighter-weight. I didn't like using xerces, from the apache project, partly because I find the name irritating, and partly because I was doing some non-standard things with it and ran into problems with version changes.

There are two styles of parser interfaces. You can write a bunch of callback functions which get passed bits of the file as they're parsed and do what you want with them, or you can have the parser give you a tree structure, then have your code walk the tree when it wants to get at the data. Expat originally supported the first style of interface only, while most of the more recent parsers support both styles. I believe there's an expat project on source forge which has likely moved the interface forward a bit from the version I use.

% I've looked into xpath, but what I've read about it so far seams much % more complicated than what I need.

It depends on what you need to do with the data. If your goal is to find a specific CD, say, XPath can save you a lot of coding. If you look at all the data, it doesn't really give you anything.

% Most of the XML books I've seen
% seem to concentrate on the DTD and interoperability between HTML and % XML. I've written a DTD that I believe is right, but I haven't tested % it yet.

You don't need a DTD, although it doesn't hurt to define your data model up front, and the DTD is a reasonable way of documenting it.