[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

proposal : XML update language

Hi,

our intention is to create our own XML update language.  I have written down my
first ideas.

Andreas
<?xml version="1.0"?>
<example version="1.0">
    <adress name="Andreas">
        <fullname>Andreas Laux</fullname>
        <street>where I live</street>
        <zipcode>D-04229</zipcode>
        <town>Leipzig</town>
    </adress>
</example>

<?xml version="1.0"?>
<updates version="1.0">
    <file name="/home/andreas/xmlupdate/example/example.xml" >
        <remove select="//adress/street"
        />
    </file>
</updates>

<?xml version="1.0" encoding="UTF-8"?>
<example version="1.0">     <adress name="Andreas">
<fullname>Andreas Laux</fullname>
<zipcode>D-04229</zipcode>         <town>Leipzig</town>     </adress> </example>

Author
------
Andreas Laux <andreas@smb-tec.com>


Status
------
Working-Draft 2000-08-18


Abstract
--------
This documents describes my first idea to update a document which is stored in 
a database.


Introduction
------------
To perform an update we must have a syntax that describes the updates. All
updates are described in a XML-file. This update file is parsed by a SAXParser
to get the information and to perform all updates. The updates are specified
at some clients and the updates are performed at the server.


Design
------
At first we need a syntax, written in a file (update.xml).

<?xml version="1.0"?>
<updates version="1.0">

    <file name="the fully specified name of the file and its location">
        
        <!--removes the selected node-->
        <remove select="a XPath expression"
        />

        <!--adds into the selected node to the given location the new element
	    or attribute or pure text-->
        <add select="a XPath expression">

            <!--adds the element and its given structure-->
            <element name="a new element">
                <attribute name="a attribute of the new element" 
                           value="and its value"
                />
                <text>
                    and some text 
                </text>
            </element>

            <!--adds the given attribute--> 
            <attribute name="a new attribute" value="and its value" 
            />

            <!--adds the given text-->
            <text>
                the text to add
            </text>

        </add>
        
	<!--changes the given node-->
        <!--update select="a true XPath expression">
            the changed text
        </update-->

    </file>

    <file name="...">

        <!--updates of another file-->

    </file>
</updates>

Now we need a content handler that is analyzing the syntax and a class to
perform these updates. The file in the database is represented by a DOM tree.
So we can select all nodes by a XPath expression. The updates can
directly performed on the DOM tree.

Conclusion
----------
Just for testing I implemented the function <remove> and run a simple example.
I attached three XML files. The example.xml is the file I stored in the
ozone database. The updates.xml is the update I performed. And the result.xml
is the file I read out the database after perfoming updates. Have a look.


Feedback and contribution
-------------------------
Feedback on this paper and contribution is very welcome at ozone-users@ozone-db.org.


References
----------
    [dom] - Document Object Model Level 2 (http://www.w3.org/TR/DOM-Level-2/)