how to avoid huge XMLELEMENT

From: Martin Drautzburg <Martin.Drautzburg_at_web.de>
Date: Mon, 20 Mar 2006 21:53:53 +0100
Message-ID: <2ch3f3-epf.ln1@C20.host>

I have to produce an XML file which represents a tree, where all nodes are basically the same type. I would like to use the xmlelement and related functions, e.g. by writing a view with a column of type XML.

The file would look like this:

<node>

<node>
</node>
<node>

                <node>
                </node>


        </node>

</node>

node can have some attribute and there can be sub-elements. But this is not important. The point is: this is a SINGLE xmlelement. There can be several million nodes in total, but they will all be crammed into a single top node.

I currently do not have the data to test this, but I am a bit worried. This xmlemement can be hundreds of megabytes in size. When I run a query that returns such an element it has to be *somewhere* using up memory. Also we sometimes convert xml into a clob using extraxt.getClobval. I don't know if this complicates things.

For this reason I usually try to retrieve xml stuff in smaller chunks, but with this nested structure I do not know how to do this. Or are my worries pointless and the total amount of data never really exists but just "virtually". Received on Mon Mar 20 2006 - 14:53:53 CST