Le mercredi 26 janvier 2005 � 11:55 +0100, Erik Bruchez a �crit : > > Just a gentle reminder... despite its esoteric title, it's asking > > serious questions :-) ... > > The mail was long and looked scary ;-)
Sorry! > >>I am creating an action to reload a XML dump of my eXist database > >>and that raises several questions showing that the behaviour of the > >>pipes called from an action isn't as straightforward as I had > >>thought. > >> > >>The bizdoc example shows several cases of such pipes without output > >>that lead to database operations and the database processors seem to > >>act as sinks that pull the input of pipes even when they don't have > >>any output, such as in: > > [...] > > >>the identity processor is never executed. That's logical and the > >>result of a good optimisation since the output of the identity > >>processor is never used. > >> > >>Is the difference of behaviour only because the oxf:xmldb-delete has > >>no output while the oxf:identity processor is it a special property > >>of the oxf:xmldb-delete processor to act as a sink? > > Yes. Processors may or may not have outputs, and there is a difference > of behavior between a processor that exposes an output if that output > is never called, and a processor that does not expose an output. And the null-serializer processor is there to force the execution of a processor which output is never called? > > In the first category, your identity processor which exposes a "data" > output which is never called. > > In the second category, the xmldb-delete processor which does not > expose an output at all. > > Without this: > > 1. As you say, we wouldn't be able to optimize and all processors > should be executed. > > 2. Or, we wouldn't be able to have processors that do not have > outputs, which would prevent implementing something like > xmldb-delete. > > >>My database dumps are RDF documents in which I'd like to split each > >>top level element in a separate document in the eXist database > >>(embedded in its own rdf root element to make it snippet a valid RDF > >>document). > >> > >>The p:for-each loop seems to be the right tool to do so. > >> > >>Unfortunately, the for-each has a mandatory output that I need to > >>declare even if the branches of the for-each are oxf:xmldb-insert > >>that have no output. > > If I remember well, the for-each's id or ref attribute is > optional. You should be able to just omit it. New tests show that this is true when there is no "root" attribute (which makes sense). > >>As a result of that, the for-each loop isn't considered as a sink > >>and isn't executed when the action is called. > > Correct, this is what will happen if you do declare an id or ref > attribute. > > >>The workaround I have found for that is to create a fake result that > >>I remove with an XPointer expression before generating the output: > > [...] > > >>That's working, but this is neither elegant, readable nor (probably) > >>efficient. > >> > >>Is there a simpler way of doing that? > > You could also simply use the null-serializer processor, which just > consumes data and does nothing with it. But you shouldn't have to do > this in this particular case. OK. > >>Would it solve the problem if the id attribute of p:for-each was made > >>optional and the for-each considered as a sink when the attribute > >>missing? > > Yes :-) If it doesn't work in 2.7, it's a bug. No, the following pipeline is working fine: <p:config xmlns:p="http://www.orbeon.com/oxf/pipeline"; xmlns:oxf="http://www.orbeon.com/oxf/processors"; xmlns:xsl="http://www.w3.org/1999/XSL/Transform"; xmlns:xdb="http://orbeon.org/oxf/xml/xmldb"; xmlns:bk="http://apiculteurs.info/namespace/"; xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#";> <p:param name="instance" type="input"/> <p:processor name="oxf:pipeline"> <p:input name="config" href="read-uri.xpl"/> <p:input name="uri" href="aggregate('uri', #instance#xpointer(string(/form/files/file)))" debug="uri"/> <p:output name="data" id="document" debug="upload"/> </p:processor> <p:for-each href="#document" select="/rdf:RDF/*"> <p:processor name="oxf:xmldb-insert"> <p:input name="datasource" href="data-access/exist/datasource.xml"/> <p:input name="query"> <xdb:insert collection="/db/apiculteurs/apiculteurs"/> </p:input> <p:input name="data" href="aggregate('rdf:RDF', current())" debug="individualFile"/> </p:processor> </p:for-each> </p:config> Thanks, Eric -- Did you know it? Python has now a Relax NG (partial) implementation. http://advogato.org/proj/xvif/ ------------------------------------------------------------------------ Eric van der Vlist http://xmlfr.org http://dyomedea.com (ISO) RELAX NG ISBN:0-596-00421-4 http://oreilly.com/catalog/relax (W3C) XML Schema ISBN:0-596-00252-1 http://oreilly.com/catalog/xmlschema ------------------------------------------------------------------------ ------------------------------------------------------- This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting Tool for open source databases. Create drag-&-drop reports. Save time by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc. Download a FREE copy at http://www.intelliview.com/go/osdn_nl _______________________________________________ orbeon-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/orbeon-user
