jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From woolly <p.b...@lbs-ltd.com>
Subject Re: Using JackRabbit for XML Structures
Date Wed, 22 Aug 2007 16:20:57 GMT

After doing a few tests, it appears that each node can take up approx 1.7k.
This seems rather a lot for a single node. Is there any way of reducing the
size...?



woolly wrote:
> 
> I've decided to store the xml as jcr:data in an nt:resource node. However,
> the hierarchy, which contains thousands of nodes, is still huge when
> stored on the file system (or in a database). Is there a way of reducing
> the amount of space these nodes take up? Perhaps using a different
> nodetype? At the moment they all just go in as nt:unstructured.
> 
> Thanks for you comments.
> 
> 
> Dan Connelly-3 wrote:
>> 
>> Phil:
>> 
>> Try placing your XML document (as jcr:data) in an nt:resource node.
>> 
>> "Import" assumes you want to have the JCR to access separate elements of 
>> XML directly, as JCR nodes.   If, on the other hand, you only want to 
>> fetch the entire XML (as a single document), then nt:resource is a more 
>> economical solution.   You can stream the docuemnt and use XPath
>> to isolate the elements of interest.
>> 
>> BTW, JBossESB bundles Jackrabbit as its "Message Store".   If your 
>> application is running on JBoss, upgrading to JBossESB just for its 
>> Message Store might be an option for you, assuming you do not need a 
>> full-blown Content Management System.  JBossESB has less functionality 
>> but gives you a useful, supported, higher-level api for xml storage.   
>> 
>>        -- Dan Connelly
>> 
>> 
>> woolly wrote:
>>> Hi all,
>>>
>>> I'm trying to use JackRabbit to maintain XML documents. The problem is
>>> that
>>> these XML documents can get quite large. When I try to import an 8mb
>>> document, it becomes 180mb in the repository (storing it just on the
>>> filesystem), and it takes about 5 minutes to go in. 
>>> I'm just using...
>>> session.importXML(node.getPath(), fis,
>>> ImportUUIDBehavior.IMPORT_UUID_CREATE_NEW);
>>> ...to import the xml, where the node is of type nt:unstructured.
>>>
>>> Is there a configuration or method that I should be using in order to
>>> let
>>> jackrabbit know that i'm dealing with xml?
>>>
>>> Thanks for any help.
>>>
>>> Phil.
>>>   
>> 
>> 
>> 
> 
> 

-- 
View this message in context: http://www.nabble.com/Using-JackRabbit-for-XML-Structures-tf4284970.html#a12278064
Sent from the Jackrabbit - Users mailing list archive at Nabble.com.


Mime
View raw message