wxcode-users

[wxCode-users] wxXml2

[Ildar <ild...@gm...>]

hi all,

I'm new here and might miss the page that describes what I need. Please, send 
me the link to the page this case.


I'm playing with the following XML.

<?xml version="1.0" encoding="utf-8"?>
<root>
  <level1>
    <level2-1>hello</level2-1>
  </level1>
</root>

I need to get the content of /root/level1/level2-1.
My code is as follows:

wxXml2Document doc;
wxString err;

doc.Load(wxT("****.xml"), &err);

if (doc.IsOk())
{
    wxXml2Node node = doc.GetRoot();
    wxString s = node.Find(wxT("level2-1")).GetName();
    wxString s1 = node.Find(wxT("level2-1")).GetContent();
    wxString s2 = node.Find(wxT("level2-1")).GetFirstChild().GetContent();
}

the results are:

s == "level2-1" // the same as expected
s1 == "" // I expected to get "hello"
s2 == "hello" // ?


The way I'm accessing the value is quite weird and doesn't seem to be safe...

Is there an easier way?

Any links/examples/ideas are welcome.



Thank you in advance.
--
Ildar

Re: [wxCode-users] wxXml2

[Francesco M. <f18...@ya...>]

Ildar ha scritto:
> hi all,
> 
> I'm new here and might miss the page that describes what I need. Please, send 
> me the link to the page this case.
> 
> 
> I'm playing with the following XML.
> 
> <?xml version="1.0" encoding="utf-8"?>
> <root>
>   <level1>
>     <level2-1>hello</level2-1>
>   </level1>
> </root>
> 
> I need to get the content of /root/level1/level2-1.
> My code is as follows:
> 
> wxXml2Document doc;
> wxString err;
> 
> doc.Load(wxT("****.xml"), &err);
> 
> if (doc.IsOk())
> {
>     wxXml2Node node = doc.GetRoot();
>     wxString s = node.Find(wxT("level2-1")).GetName();
>     wxString s1 = node.Find(wxT("level2-1")).GetContent();
>     wxString s2 = node.Find(wxT("level2-1")).GetFirstChild().GetContent();
> }
> 
> the results are:
> 
> s == "level2-1" // the same as expected
> s1 == "" // I expected to get "hello"
> s2 == "hello" // ?
yes, sure. no bugs.
This is really a FAQ and I should add it somewhere.

The results you get are because your XML tree is parsed as:

wxXML_ELEMENT_NODE with name "root" and content=""
|- wxXML_ELEMENT_NODE with name "level1" and content=""
    |- wxXML_ELEMENT_NODE with name "level2-1" and content=""
       |- wxXML_TEXT_NODE with name "" and content="hello"

Thus, if the node is an element node, you shouldn't look at its contents 
but rather at its children.
If the node is a text node, the name is empty (or maybe fixed to "text", 
I don't remember) and you should just look at his contents.

I've added in CVS the GetNodeContent() function which simplifies the 
process (see below).


> The way I'm accessing the value is quite weird and doesn't seem to be safe...
 >
 > Is there an easier way?

wxXml2Node child = doc.GetRoot();
while (child != wxXml2EmptyNode) {

     if (child.GetName() == wxT("level1")) {

         // process text enclosed by <level1>
         wxString content = child->GetNodeContent();

     } else if (child->GetName() == wxT("tag2")) {

         // process tag2 ...
     }
     else
     {
         // unknown tag?
     }

     child = child.GetNext();
}

this pattern allows you to scan each node of the XML and e.g. catch 
structural errors (e.g. you can check for unknown tags).


Francesco

Re: [wxCode-users] wxXml2

[Ildar <ild...@gm...>]

Francesco,

Thank you, I got the idea.

However, 2 more questions:

#1

wxString s3 = node.Find(wxT("level2-1")).GetFirstChild().GetName();

result:
s3 == "text" // as you mentioned

but 

wxString s4 = node.Find(wxT("level2-1")).Get(wxT("text"));

result:
s4 == "" 

shouldn't it return "hello"?

#2

is the following possible?

     |- wxXML_ELEMENT_NODE with name "level2-1" and content=""
        |- wxXML_TEXT_NODE with name "" and content="hello"
        |- wxXML_TEXT_NODE with name "" and content="bye"

wxXML_ELEMENT_NODE - contains 2 text nodes.

Re: [wxCode-users] wxXml2

[Francesco M. <f18...@ya...>]

Ildar ha scritto:
> Francesco,
> 
> Thank you, I got the idea.
> 
> However, 2 more questions:
> 
> #1
> 
> wxString s3 = node.Find(wxT("level2-1")).GetFirstChild().GetName();
> 
> result:
> s3 == "text" // as you mentioned
> 
> but 
> 
> wxString s4 = node.Find(wxT("level2-1")).Get(wxT("text"));
> 
> result:
> s4 == "" 
wxXml2Node::Get returns a wxXml2Node, not a wxString. I don't know why 
the compiler accepts the assignment of a wxXml2Node to a wxString...
couldn't you trace it with a debugger and see what happens?


> #2
> 
> is the following possible?
> 
>      |- wxXML_ELEMENT_NODE with name "level2-1" and content=""
>         |- wxXML_TEXT_NODE with name "" and content="hello"
>         |- wxXML_TEXT_NODE with name "" and content="bye"
> 
> wxXML_ELEMENT_NODE - contains 2 text nodes.
sure; it's possible. IIRC there are some rules which makes libxml2 split 
a text into two text nodes but I don't remember them right now.

However consider also the case of this XML fragment:

    <node>
      hello <b>XML</b> world!
    </node>

the "node" node has two text nodes interleaved by an element node (with 
name="b" and with a single text node as child).


Francesco

Re: [wxCode-users] wxXml2

[Ildar <ild...@gm...>]

> > However, 2 more questions:
> > 
> > #1
> > 
> > wxString s3 = node.Find(wxT("level2-1")).GetFirstChild().GetName();
> > 
> > result:
> > s3 == "text" // as you mentioned
> > 
> > but 
> > 
> > wxString s4 = node.Find(wxT("level2-1")).Get(wxT("text"));
> > 
> > result:
> > s4 == "" 
> wxXml2Node::Get returns a wxXml2Node, not a wxString. I don't know why 
> the compiler accepts the assignment of a wxXml2Node to a wxString...
> couldn't you trace it with a debugger and see what happens?

Sorry, the code is as follows

wxString s4 = node.Find(wxT("level2-1")).Get(wxT("text")).GetContent();

Btw, 

wxString s5 = node.Find(wxT("level2-1")).Get(wxT("text")).GetName();

result:
s5 == "" 

Re: [wxCode-users] wxXml2

[Francesco M. <f18...@ya...>]

Ildar ha scritto:
>>> However, 2 more questions:
>>>
>>> #1
>>>
>>> wxString s3 = node.Find(wxT("level2-1")).GetFirstChild().GetName();
>>>
>>> result:
>>> s3 == "text" // as you mentioned
>>>
>>> but 
>>>
>>> wxString s4 = node.Find(wxT("level2-1")).Get(wxT("text"));
>>>
>>> result:
>>> s4 == "" 
>> wxXml2Node::Get returns a wxXml2Node, not a wxString. I don't know why 
>> the compiler accepts the assignment of a wxXml2Node to a wxString...
>> couldn't you trace it with a debugger and see what happens?
> 
> Sorry, the code is as follows
> 
> wxString s4 = node.Find(wxT("level2-1")).Get(wxT("text")).GetContent();
> 
> Btw, 
> 
> wxString s5 = node.Find(wxT("level2-1")).Get(wxT("text")).GetName();
It looks like Get() does not find that node. Can you trace it with a 
debugger ?

It should be easy to understand wxXml2 source code and find what's 
wrong... (even if I suspect that's not really a bug).

Thanks,
Francesco

Re: [wxCode-users] wxXml2

[Ildar <ild...@gm...>]

and one more ;)

is there a way to get a count of nodes?

e.g. 

- Find(...) allows to get the N-th node which meets the condition, but what's 
the number of nodes which meet the condition?

- or how to get the number of wxXML_ELEMENT_NODE subnodes of wxXML_ELEMENT_NODE 
node?


I understand I can enumerate the nodes and count them, is it the only possible 
solution?

Re: [wxCode-users] wxXml2

[Francesco M. <f18...@ya...>]

Ildar ha scritto:
> and one more ;)
> 
> is there a way to get a count of nodes?
currently no; that's because wxXml2 API derives from wxXml API because 
when I wrote it I wanted to port some wxXml code to a newer API with DTD 
validation and thus I made wxXml2 API mostly compatible with wxXml one, 
which has no GetChildCount() methods.

However I agree this would be useful to have. Patches welcome :D

> e.g. 
> 
> - Find(...) allows to get the N-th node which meets the condition, but what's 
> the number of nodes which meet the condition?
we could add a new GetChildrenCount(const wxXmlNode &tofind, bool bNS = 
TRUE, bool recurse = TRUE) function to return it


> - or how to get the number of wxXML_ELEMENT_NODE subnodes of wxXML_ELEMENT_NODE 
> node?
I would add a GetChildrenCount() const function for this.


> I understand I can enumerate the nodes and count them, is it the only possible 
> solution?
it highly depends on how to you want to organize your parser.
For me, the pattern I copied & pasted in previous mail has always been a 
winning pattern ;)

I.e. you scan all the XML document, handle unknown tags in the else{} 
part, and call subroutines to load other nodes in other elseif{} branches.

It's a lot of time I don't look at XML Schemas, SAX parser, etc.
You can get an idea of the "low-level" libxml2 API looking at 
http://xmlsoft.org/html/index.html

Francesco

Re: [wxCode-users] wxXml2

[Armel A. <ass...@wa...>]

>> the results are:
>>
>> s == "level2-1" // the same as expected
>> s1 == "" // I expected to get "hello"
>> s2 == "hello" // ?
> yes, sure. no bugs.
> This is really a FAQ and I should add it somewhere.
>
> The results you get are because your XML tree is parsed as:
nonetheless, this is not really XSL-like (which Ildar expected). It could be 
cool to have a GetAllText ( ) all the simple text of all sub nodes 
concatenated.

Armel

Re: [wxCode-users] wxXml2

[Francesco M. <f18...@ya...>]

Armel Asselin ha scritto:
>>> the results are:
>>>
>>> s == "level2-1" // the same as expected
>>> s1 == "" // I expected to get "hello"
>>> s2 == "hello" // ?
>> yes, sure. no bugs.
>> This is really a FAQ and I should add it somewhere.
>>
>> The results you get are because your XML tree is parsed as:
> nonetheless, this is not really XSL-like (which Ildar expected). It could be 
> cool to have a GetAllText ( ) all the simple text of all sub nodes 
> concatenated.
I just added to CVS the GetNodeContent() functions which does exactly that.

Francesco

Re: [wxCode-users] wxXml2

[Armel A. <ass...@wa...>]

>> nonetheless, this is not really XSL-like (which Ildar expected). It could 
>> be
>> cool to have a GetAllText ( ) all the simple text of all sub nodes
>> concatenated.
> I just added to CVS the GetNodeContent() functions which does exactly 
> that.
cool :)

Armel 

Re: [wxCode-users] wxXml2

[Ildar <ild...@gm...>]

> I just added to CVS the GetNodeContent() functions which does exactly that.

what do you think of adding "recursive" param to the function?

e.g.
GetNodeContent(true) would work as it works now - it would return a 
concatenation string of all sub-nodes.

GetNodeContent(false) would return a concatenation string of wxXML_TEXT_NODE 
nodes which belong to current wxXML_ELEMENT_NODE