[
https://issues.apache.org/jira/browse/CAMEL-3998?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13051697#comment-13051697
]
Christian Müller commented on CAMEL-3998:
-----------------------------------------
In one of our projects, we need the ability to split a large XML file (3 - 10
GB) into its individual records. Because XPath reads the entire document into
memory, it's not an option for us. For the time being, we will develop our own
custom Splitter which use internally StAX to stream the file and return an
Iterator. For each call of hasNext()/next() on the Iterator interface, we will
read and return the next record.
I think we will share this solution so that in Camel 2.9.0, the user should be
able to do something like (in the same way as the xpath support):
{code:java}
from("direct:start")
.split(stax("//foo/bar"))
.convertBodyTo(String.class) // may be not necessary, depending on the stax
response
.to("mock:result")
{code}
The following should also be possible:
{code:java}
StaxBuilder staxBuilder = new StaxBuilder("//foo/bar");
from("direct:start")
.split(staxBuilder)
.convertBodyTo(String.class) // may be not necessary, depending on the stax
response
.to("mock:result")
{code}
To support namespaces, the following should be implemented:
{code:java}
Namespaces ns = new Namespaces("c", "http://acme.com/cheese";);
from("direct:start")
.split(stax("//foo/bar", ns))
.convertBodyTo(String.class) // may be not necessary, depending on the stax
response
.to("mock:result")
{code}
and also:
{code:java}
Namespaces ns = new Namespaces("c", "http://acme.com/cheese";);
StaxBuilder staxBuilder = new StaxBuilder("//foo/bar");
staxBuilder.setNamespaces(ns.getNamespaces());
from("direct:start")
.split(staxBuilder)
.convertBodyTo(String.class) // may be not necessary, depending on the stax
response
.to("mock:result")
{code}
What do you think?
> improve stax/jax components
> ---------------------------
>
> Key: CAMEL-3998
> URL: https://issues.apache.org/jira/browse/CAMEL-3998
> Project: Camel
> Issue Type: Improvement
> Components: camel-core
> Reporter: Romain Manni-Bucau
> Priority: Minor
>
> it could be pretty cool to have an iterator to be able to iterate over a
> filtered tree of a xml document avoiding to load all the document in memory.
> it can be nice to do it loading sub objects using jaxb.
> i did a version here:
> http://code.google.com/p/rmannibucau/source/browse/camel/camel-stax/src/main/java/org/apache/camel/stax/expression/StAXJAXBIteratorExpression.java
> (i added some converters to help too).
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
