jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marcel Reutegger <marcel.reuteg...@gmx.net>
Subject Re: Indexing XML stored as property in nt:unstructured
Date Wed, 22 Aug 2007 09:51:05 GMT
Daniel Br┬čuen wrote:
> As mentioned before, I need to index XML-data that ist stored in a 
> property of a nt:unstructured node.
> If I understand the whole indexing-procedure correctly I need to define 
> the correct mimeType of the data, so that the correct textExtractor is 
> used (XMLTextExtractor in this case).
> Can I just add two properties "jcr:mimeType" (="text/xml") and 
> "jcr:data" (="<xml>my xml</xml>") to an unstructured node to get this 
> working?

this might work, but I haven't tested it. the text extractors were built to work 
on nt:resource nodes. but they don't actually check the node types, they're just 
looking for the properties as defined in nt:resource.

> I know I can configure which properties are to be indexed in 
> repository.xml/workspace.xml, but I can not attach a special extractor 
> to the property there...
> 
> Or do I have to make use of nt:file and nt:resource (which I would need 
> to extend then)?

yes, I recommend you use the predefined node type. using predefined node types 
always opens up your content to existing tools that know how to handle the 
respective content.

regards
  marcel

Mime
View raw message