need to parse XML file without root elem

[akira]

Hello,
I need to parse this XML file or XML stream . Is it possible to parse
it with DOM reading each DataContainer one by one.

<DataContaine r>
<name> </name>
<birthDate> </birthDate>
<personnalInfos >
<personnalInfo_ entries> </personnalnfo_en tries>
<personnalInfo_ elem> </personnalnfo_el em>
<personnalInfo_ elem> </personnalnfo_el em>
</personnalInfos>
</DataContainer>
<DataContaine r>
<name> </name>
<birthDate> </birthDate>
<personnalInfos >
<personnalInfo_ entries> </personnalnfo_en tries>
<personnalInfo_ elem> </personnalnfo_el em>
<personnalInfo_ elem> </personnalnfo_el em>
</personnalInfos>
</DataContainer>

Cheers.

[Felix Engelhardt]

> I need to parse this XML file or XML stream . Is it possible to parse

it with DOM reading each DataContainer one by one.
No, DOM documents are required to have exactly one root node, not several.
<DataContaine r>
<name> </name>
<birthDate> </birthDate>
<personnalInfos >
<personnalInfo_ entries> </personnalnfo_en tries>
<personnalInfo_ elem> </personnalnfo_el em>
<personnalInfo_ elem> </personnalnfo_el em>
</personnalInfos>
</DataContainer>
<DataContaine r>
<name> </name>
<birthDate> </birthDate>
<personnalInfos >
<personnalInfo_ entries> </personnalnfo_en tries>
<personnalInfo_ elem> </personnalnfo_el em>
<personnalInfo_ elem> </personnalnfo_el em>
</personnalInfos>
</DataContainer>

[akira]

What is the solution then? I can not have a root elem solution as I can
have a hudge file.

parse with sax? Any idea?
Thx

[Richard Tobin]

In article <11************ **********@j33g 2000cwa.googleg roups.com>,
akira <ak*******@yaho o.fr> wrote:

What is the solution then? I can not have a root elem solution as I can
have a hudge file.

You won't be able to read it with an XML parser if you don't put a
root element around it. I'm not sure why the size of the file should
make it harder to add one.

One way to get a root element is to create a short file with an entity
reference pointing to the "real" data.

<!DOCTYPE dummy [
<!ENTITY data SYSTEM "http://wherever-my-data-is">
]>
<dummy>
&data;
</dummy>

--- Richard

[Felix Engelhardt]

akira wrote:

What is the solution then? I can not have a root elem solution as I can
have a hudge file.

parse with sax? Any idea?

Afaik, SAX is more suitable for large documents than DOM. I don't know if it
supports multiple root nodes. With SAX, afaik, it's not possible to modify
or create documents, you can only parse them and then they are static.
You might also have a look at JDOM or DOM4J, though I don't know much about
these. Just to give you some keywords.

I don't yet understand why having one root node instead of several is a
problem with large documents.

If the reason is that you want to split your data across multiple files, you
could still have a single root node in each one, and then use DOM to
assemble them after you have loaded them. Or (which is a bit dirty) you just
concatenate the files as strings, write "<my-root-node>" at the beginning
and "</my-root-node>" at the end and then parse this string using DOM.

Hope this helps.

Regards,

Felix

[Jürgen Kahrs]

Felix Engelhardt wrote:

I need to parse this XML file or XML stream . Is it possible to parse
it with DOM reading each DataContainer one by one.

No, DOM documents are required to have exactly one root node, not several.

That's true. Especially because well-formed XML files
are allowed to have one root only.

<DataContaine r>
<name> </name>
<birthDate> </birthDate>
<personnalInfos >
<personnalInfo_ entries> </personnalnfo_en tries>
<personnalInfo_ elem> </personnalnfo_el em>
<personnalInfo_ elem> </personnalnfo_el em>
</personnalInfos>
</DataContainer>
<DataContaine r>
<name> </name>
<birthDate> </birthDate>
<personnalInfos >
<personnalInfo_ entries> </personnalnfo_en tries>
<personnalInfo_ elem> </personnalnfo_el em>
<personnalInfo_ elem> </personnalnfo_el em>
</personnalInfos>
</DataContainer>

The subject of parsing XML data with multiple roots
comes up about once each month. If you are bound to
usual XSL processors or DOM libraries, then there's
no way out for you. If you are free to choose a tool
with a SAX-like processing style, then you might be
lucky. For example, the XMLgawk extension of the GNU
Awk interpreter allows you to parse XML data with
multiple roots:

http://home.vrweb.de/~juergen.kahrs/...wk-interpreter

One of the authors (Andrew Schorr) processes GigaBytes
of such data and stores it in a database.

[akira]

Thanks a lot

[Joseph Kesselman]

Felix Engelhardt wrote:

No, DOM documents are required to have exactly one root node, not several.

Because a well-formed XML document is required to have exactly one
top-level element, not several. The file you've shown us is not an XML
document, though it may be an XML External Entity.

DOM DocumentFragmen ts can have several top-level elements. But I don't
know of an off-the-shelf parser that will read an External Entity
directly into a DocumentFragmen t.

One solution is to set up a bit of additional code which wraps a
top-level element around the data as it's read in -- for example, set up
a SAXFilter which generates an extra startElement after the
startDocument event and an extra endElement before the endDocument
event. That does require that the rest of your code be aware this dummy
element was introduced, but that's usually straightforward ... or you can
move these trees into a DocumentFragmen t, if you're working with DOMs.

Saving the data back out may be complicated since, like parsers,
serializers may or may not assume they're intended to process
well-formed XML document. Generally, though, they're more forgiving and
you can get away with using the standard serializer; it's just your
responsibility not to write out the dummy wrapper element.

Good luck.

--
Joe Kesselman / Beware the fury of a patient man. -- John Dryden