XML to XML through XSLT

ratzko · December 20th, 2007, 04:49 AM

I have a source and a resultant xml.Can someone provide me with an xslt for this scenario.

Source:

Code:

<text>With the Linux installation package, you receive the Linux
    Installer application that installs the software and databases. <p />Best installation
    instructions are available from HOWTO site
    (http://www.google.com/howto/). If possible, we suggest that you use
    one of these: 
        <list>
            <li><a href="http://www.google.com/howto/install_linux.html">HOWTO Install Linux</a></li>
            <li><a href="http://www.yandex.ru/howto/install_win.html">HOWTO Install Windows</a></li>
        </list>
    Linux can be extended with various plugins. If you are in need of extended functionality, please check
    <a href="http://www.google.com/">www.google.com/</a>.</text>
    </text>

Resultant:

Code:

<root>
    <text:p>
    With the Linux installation package, you receive the Linux
    Installer application that installs the software and databases. 
    </text:p>
    <text:p/>
    <text:p>
    Best installation
    instructions are available from HOWTO site
    (http://www.google.com/howto/). If possible, we suggest that you use
    one of these: 
    </text:p>
    <text:list>
        <text:list-item>
            <text:p>
                <text:a xlink:href="http://www.google.com/howto/install_linux.html">
                    <text:span>HOWTO Install Linux</text:span>
                </text:a>
            </text:p>
        </text:list-item>
        <text:list-item>
            <text:p>
                <text:a xlink:href="http://www.yandex.ru/howto/install_win.html">
                    <text:span>HOWTO Install Windows</text:span>
                </text:a>
            </text:p>
        </text:list-item>
    </text:list>
    <text:p>
    Linux can be extended with various plugins. If you are in need of extended functionality, please check
        <text:a xlink:href="http://www.google.com.ru/">
            <text:span>www.google.com</text:span>
        </text:a>
    </text:p>
</root>

samjudson · December 20th, 2007, 05:14 AM

Generally speaking people wont just write full solutions for you because you ask nicely - that is what most people get paid to do.

Your best bet is to give it a go yourself, and if you get stuck come back and ask specific question. You might like to start with a simple web tutorial such as that provided by the W3C Schools web site. http://www.w3schools.com/xsl/default.asp

/- Sam Judson : Wrox Technical Editor -/

mhkay · December 20th, 2007, 06:09 AM

You might be lucky and find someone who's in the Chirstmas spirit, but my own approach on this list is that I'm happy to help people who have a specific question or a specific difficulty, but I don't provide a free programming service.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

ratzko · December 20th, 2007, 07:11 AM

OK. I understood.

I have wrote next code:

Code:

<xsl:template match="text">
    <xsl:apply-templates/>
</xsl:template>

<xsl:template match="text()">
    <text:p>
        <xsl:value-of select="." />
    <text:p>
</xsl:template>

<xsl:template match="p">
    <text:p/>
</xsl:template>

<xsl:template match="a">
    <text:a>
        <xsl:attribute name="xlink:href">
        <xsl:value-of select="@href" />
    </xsl:attribute>
       <text:span text:style-name="T12"><xsl:apply-templates/></text:span>
    </text:a>
</xsl:template>


<xsl:template match="list">
    <text:list>
        <xsl:apply-templates/>
    </text:list>
</xsl:template>

<xsl:template match="li">
    <text:list-item>
         <text:p text:style-name="P19">
             <xsl:apply-templates/>
         </text:p>         
     </text:list-item>
</xsl:template>

But this is not exactly what I wont.
How to put <text:a> with text into the tag <text:p > without <text:list>?

mhkay · December 20th, 2007, 07:48 AM

I had to fix a few well-formedness errors in both your XML and your XSLT code.

I think the main difficulty here is knowing when to wrap a text node in <text:p> and when not. It's difficult to work out what all the rules are from such a small sample of text. It seems to me, for example, that you want the <text:p> wrapper for text that appears inside a <text> element, but not for text that appears inside an <a> element. That suggests you need different rules for the two cases:

<xsl:template match="text/text()">
<text:p><xsl:value-of select="."/></text:p>
</xsl:template>

<xsl:template match="a/text()">
<xsl:value-of select="."/>
</xsl:template>

However, I'm unable to generalize from this as to what the default rule for text nodes should be, or the rule for text nodes in other contexts - there's not enough data to go on.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

ratzko · December 20th, 2007, 08:54 AM

I have no difficult's with text wrapping. I don't understand how can I wrap some nodes with they following or preceding text in <text:p>:

Code:

<text:p>
    Linux can be extended with various plugins. If you are in need of     extended functionality, please check
        <text:a xlink:href="http://www.google.com.ru/">
            <text:span>www.google.com</text:span>
        </text:a>
    </text:p>

and some don't wrap in <text:p>:

Code:

<text:list>
        <text:list-item>
            <text:p>
                <text:a xlink:href="http://www.google.com/howto/install_linux.html">
                    <text:span>HOWTO Install Linux</text:span>
                </text:a>
            </text:p>
        </text:list-item>
    </text:list>

samjudson · December 20th, 2007, 09:36 AM

You problem is that your original XML isn't designed very well to try to get the output you want.

Basically you have an XML structure that looks something like this:

- <text>
  - text() 'With the Linux...'
  - <p />
  - text() 'Best installation...'
  - <list>
    - <li>
      - <a>
        - text() 'HOWTO Install Linux'
    - <li>
      - <a>
        - text() 'HOWTO Install Windows'
  - text() 'Linux can be extended...'
  - <a>
    - text() 'http://www.google.com'
  - text() '.'

As you can see inside your list the text() nodes (not to be confused with your <text> elements) are nicely inside the <a> element, however at the end the <a> element is a sibling to the two text() nodes.

The outcome you are trying to get is to wrap the two text() nodes and the <a> element in one <text:p> element - which isn't going to be easy. Off the top of my head I can't think of a simple or elegant solution at all.

/- Sam Judson : Wrox Technical Editor -/

ratzko · December 21st, 2007, 05:24 AM

I know that XML isn't designed very well for my need, but I can't change this.

Can I close tag <text> when meet tag <list> and after list close that?

Thanx for your help!

mhkay · December 21st, 2007, 05:43 AM

XSLT doesn't write tags to an output file: it writes nodes to a tree. Later, the serializer takes the tree and turns an element node into a start-tag/end-tag pair, but in your stylesheet you have to think of instructions writing nodes - and you can't write half a node.

This means that when you have a complex problem it's a good idea sometimes to sketch the input and output trees on paper *as trees* and then analyze the correspondence between nodes in the input and nodes in the output. The difficulties tend to arise when one node in the output corresponds to a group of nodes in the input. That's known as a grouping problem, and there's a range of techniques for dealing with them.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference