Parsing XML file into C structs

I have a massive XML file that I need to parse into C.

Code:

  <node id="4700-Z$0" x="322" y="-505" z="0.0" /> 
  <node id="4701" x="356" y="-454" z="0.0" /> 
  <node id="4702-H" x="402" y="-456" z="0.0" /> 
  <node id="4704" x="400" y="-428" z="0.0" /> 
  <node id="4705" x="455" y="-426" z="0.0" />

thats just a small sample code of the XML. So i want to extract the Node, x,y,z.

Into something like..

Code:

typedef struct{
              char node[10];
               int x; int y; int z;
} graphJ;

I tried looking at the libxml2 but can't get it to build, and I don't think I need it for something somewhat simple for this... I think?

Sounds like mini-xml is what you need Mini-XML

I've use it before, it's great.

this would be great except im using visual studios

It doesn't say anywhere that MS Visual Studio DOESN'T work - it says that it WORKS with gcc. I'd give it a go.

--
Mats

Quick and dirty...

Code:

	char *tag_id = "id=\"";
	char *tag_x = "x=\"";
	char *tag_y = "y=\"";
	char *tag_z = "z=\"";

	typedef struct{
              char node[10];
               int x; int y; int z;
	} graphJ_def;

	graphJ_def graphJ[10];
	int n = 0;
	char *p;

	// for every line input...

	sscanf(strstr(tests, tag_id) + strlen(tag_id), "%s", graphJ[n].node);
	graphJ[n].node[strlen(graphJ[n].node) - 1] = '\0';	// kill traling quote

	sscanf(strstr(tests, tag_x) + strlen(tag_x), "%d", &graphJ[n].x);

	sscanf(strstr(tests, tag_y) + strlen(tag_y), "%d", &graphJ[n].y);

	sscanf(strstr(tests, tag_z) + strlen(tag_z), "%d", &graphJ[n].z);

	n++;

Originally Posted by nonoob

Quick and dirty...

Yeah, but it's got so much potential for disaster. If any of those strstr()s fail, if there's some oddity like an entity, the XML isn't ascii, other tags, the id attribute contains an x, y or z, etc. (What's the char *p for?)
And z appears to be a float/double, btw.
But, like you said, quick and dirty.

I'd take the time to learn an XML library. If you've not compiled many libraries before, this is a good opportunity. (I've compiled libxml2 on Linux & Windows, granted, I used gcc both times... perhaps they have a binary download?) You'll run into full blown XML parsing sooner or later.
You could XSLT it into an easier format, perhaps too. (I'm learning XSLT at the moment... so it's at the top of my head...)

I like the xerces-c library for xml

If all of your data follows such a strict layout -- i.e., one tag per line etc -- then you could use some scripting to convert the data into a more easily parseable form. Just a thought. I wouldn't do this if your program will be run often in the future because it's not very robust, but a simple script like this could make your life a lot easier in C:

Code:

$ cat input
  <node id="4700-Z$0" x="322" y="-505" z="0.0" />
  <node id="4701" x="356" y="-454" z="0.0" />
  <node id="4702-H" x="402" y="-456" z="0.0" />
  <node id="4704" x="400" y="-428" z="0.0" />
  <node id="4705" x="455" y="-426" z="0.0" />
$ perl -ne 'while(/\s(?:id|[xyz])="([^"]+)"/g) { print "$1 " } print "\n"' input
4700-Z$0 322 -505 0.0
4701 356 -454 0.0
4702-H 402 -456 0.0
4704 400 -428 0.0
4705 455 -426 0.0
$

That's in the spirit of even more quick and dirty solutions. If this program is actually going to be used more than once, for sure look into libmxml. I've used it before too (thanks to zacs7!) and would highly recommend it. libexpat is good too.