Mailing-List: contact users-help@jackrabbit.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@jackrabbit.apache.org
Received-SPF: pass (herse.apache.org: domain of alessandro.bologna@gmail.com
 designates 209.85.132.245 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=beta;
        h=received:mime-version:in-reply-to:references:content-type:message-id:content-transfer-encoding:from:subject:date:to:x-mailer;
        b=cNXIT9F4xrghgKE0woRLZ9a1wCxPPlK7yJXqQH3jtE79NgUN7XMzhUrtNnA1Y0XoJP4Oj5LPCkRok2li43QAmFedJHVxkoqdMvZg+fIMfxSd1aG3dBKNhyVJTXEI5C1UoioRc7NEqJzfod+m0ZXFbPIjx7WClNlLaouuoTHmL1s=
Mime-Version: 1.0 (Apple Message framework v752.3)
In-Reply-To: <90a8d1c00706130522k42f55f3fn3ee8a32006a672e5@mail.gmail.com>
References: <FF73C1ED-1879-4840-90F4-75C7E5893A0A@gmail.com>
 <EC213738EDD1C348BBE6012077FB746B07789E@asterion.legisway.com>
 <90a8d1c00706130522k42f55f3fn3ee8a32006a672e5@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; delsp=yes; format=flowed
Message-Id: <1DBE6C14-8BEF-4E10-989F-128AFD893F7B@gmail.com>
Content-Transfer-Encoding: quoted-printable
From: Alessandro Bologna <alessandro.bologna@gmail.com>
Subject: Re: Importing and Exporting XML
Date: Wed, 13 Jun 2007 08:39:03 -0400
To: users@jackrabbit.apache.org

Stephan,

I agree that DocumentView will not allow proper roundtripping (when =20
the source is a JCR model). At the same time, when the source is an =20
XML document, DocumentView does (and SystemView does not), and this =20
was the case in question.

To address Frederic point of view, I think that when it comes to =20
multivalued properties (which do not really happen often in XML, so =20
we are dealing with roundtripping from a JCR model), one approach is =20
to use SystemView (and then maybe transform it in something more =20
human readable), or to use a SAX handler and the visitor pattern for =20
a customized output (as in the DocumentViewExportVisitor class).
Of course, I realize that there can be use cases when it is not =20
practical, and then probably there's no one solution-fits-all.

Alessandro


On Jun 13, 2007, at 8:22 AM, Stefan Guggisberg wrote:

> On 6/13/07, Fr=E9d=E9ric Esnault <fesn@legisway.com> wrote:
>> Of course, in wolly's case, it doesn't really matter, but =20
>> jackrabbit is not supposed to be  a "one use product", but a =20
>> generic JCR. And I just tried my own use case, which is not a very =20=

>> odd case, where I have some contracts referencing contractors by =20
>> reference. And with a document view, I loose these references, the =20=

>> contract managers also multivalued, and so on.
>>
>> I don't know how you export references, because I can't.
>
> the answer is simple: use system view export.
>
> see http://issues.apache.org/jira/browse/JCR-325 for a
> discussion of multi-valued property representation in the
> document view xml mapping.
>
> please see also section 6.4.2.5 in the JSR 170 spec.
>
> cheers
> stefan
>
>
>
>>
>> Fr=E9d=E9ric Esnault
>>
>> -----Message d'origine-----
>> De: Alessandro Bologna [mailto:alessandro.bologna@gmail.com]
>> Envoy=E9: mercredi 13 juin 2007 13:55
>> =C0: users@jackrabbit.apache.org
>> Objet: Re: Importing and Exporting XML
>>
>> Frederic,
>>
>> Some multivalued properties are handled correctly (say for instance
>> multivalued references are exported as space separated lists of
>> references in an attribute), and in some other cases it's still
>> possible to do an encoding of the values, so that they can be
>> represented inside an XML attribute. But I agree that in some cases
>> it's really not practical (say binary properties...) because the XML
>> that would result would probaby be not very much readable.
>>
>> But in Wolly's  (or Phil's?) case, the problem is to round-trip XML,
>> which does not have multi-valued properties to start with (so it
>> doesn't really matter, does it?)
>>
>> Alessandro.
>>
>>
>> On Jun 13, 2007, at 7:43 AM, Fr=E9d=E9ric Esnault wrote:
>>
>> > What about the multivalued properties not included in the export?
>> > The exported document, even if it's modified by XSL/java is still
>> > incorrect, as the multivalued properties are simply ignored.
>> >
>> > I know the JSR 170 spec leaves the right to do so, but imo it
>> > creates an incoherent export. I don't think anyone even think about
>> > the possibility that some properties are just ignored when doing an
>> > export from a CMS...
>> >
>> > Fr=E9d=E9ric Esnault - Ing=E9nieur R&D
>> >
>> >
>> > -----Message d'origine-----
>> > De : Alessandro Bologna [mailto:alessandro.bologna@gmail.com]
>> > Envoy=E9 : mercredi 13 juin 2007 13:34
>> > =C0 : users@jackrabbit.apache.org
>> > Objet : Re: Importing and Exporting XML
>> >
>> > Yes. Two ways:
>> > 1. Java way: use the DocumentViewExportVisitor and override the
>> > includeProperty() method (simply return false if the property has a
>> > namespace you don't want).
>> >
>> > 2. XML way: use an XSLT identity transformation and add rules for
>> > removing any attribute that you don't want.
>> >
>> > Alessandro
>> >
>> >
>> > On Jun 13, 2007, at 7:20 AM, woolly wrote:
>> >
>> >>
>> >> Thanks for replying...
>> >>
>> >> If I have a document like this:
>> >> <?xml version=3D"1.0" encoding=3D"UTF-8"?>
>> >> <cheeses>
>> >>      <cheese>edam</cheese>
>> >>      <cheese>lancashire</cheese>
>> >>      <cheese>cheshire</cheese>
>> >>      <cheese>stilton</cheese>
>> >> </cheeses>
>> >>
>> >> ...and I import it using:
>> >> fis =3D new FileInputStream(inputFile);
>> >> session.importXML(node.getPath(), fis,
>> >> ImportUUIDBehavior.IMPORT_UUID_CREATE_NEW);
>> >> fis.close();
>> >>
>> >> ...when I export it using:
>> >> out =3D new FileOutputStream(outputFile);
>> >> session.exportDocumentView(node.getPath(), out, true, false);
>> >>
>> >> ...I get:
>> >>
>> >> <?xml version=3D"1.0" encoding=3D"UTF-8"?>
>> >> <cheeses xmlns:dc=3D"http://www.purl.org/dc/elements/1.1/"
>> >> xmlns:jcr=3D"http://www.jcp.org/jcr/1.0"
>> >> xmlns:nt=3D"http://www.jcp.org/jcr/nt/1.0"
>> >> xmlns:sv=3D"http://www.jcp.org/jcr/sv/1.0"
>> >> xmlns:rdf=3D"http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>> >> xmlns:xs=3D"http://www.w3.org/2001/XMLSchema"
>> >> xmlns:fn=3D"http://www.w3.org/2005/xpath-functions"
>> >> xmlns:fn_old=3D"http://www.w3.org/2004/10/xpath-functions"
>> >> xmlns:mix=3D"http://www.jcp.org/jcr/mix/1.0" xmlns:rep=3D"internal"
>> >> jcr:primaryType=3D"nt:unstructured">
>> >>     <cheese jcr:primaryType=3D"nt:unstructured">edam</cheese>
>> >>     <cheese jcr:primaryType=3D"nt:unstructured">lancashire</cheese>
>> >>     <cheese jcr:primaryType=3D"nt:unstructured">cheshire</cheese>
>> >>     <cheese jcr:primaryType=3D"nt:unstructured">stilton</cheese>
>> >> </cheeses>
>> >>
>> >> ...is there any way to just get the original document out?
>> >>
>> >> Thanks,
>> >>
>> >> Phil.
>> >>
>> >>
>> >> Julian Reschke wrote:
>> >>>
>> >>> woolly wrote:
>> >>>> Surely then I would want to use some form of document view,
>> >>>> then, in
>> >>>> order to
>> >>>> get back the original document I put in?
>> >>>
>> >>> Yes (contrary to what you've been told here before).
>> >>>
>> >>> There are known limitations with the document view (such as
>> >>> restrictions
>> >>> with multivalued props). But the thing you complained about was
>> >>> something else; maybe unneeded attributes in the jcr namespace?
>> >>> Could
>> >>> you be a bit more specific about that problem?
>> >>>
>> >>> Best regards, Julian
>> >>>
>> >>>
>> >>>
>> >>
>> >> --
>> >> View this message in context: http://www.nabble.com/Importing-and-
>> >> Exporting-XML-tf3908819.html#a11097246
>> >> Sent from the Jackrabbit - Users mailing list archive at =20
>> Nabble.com.
>> >>
>> >
>>
>>