SGMLXML.net A place for SGML and XML application developers.

Blog item that examines XQuery vs XSLT.



I would add to the insights here that XSLT is designed for transformations while XQuery was designed to find things that are not necessarily in the current instance you are working with but also not necessarily in a CMS or database.



An additional article from Michael Kay on the same subject.

Getting lots of questions about XPath on the #XML IRC channel so figured I'd write a quick cheat sheet. Sometimes the hardest part of learning something is knowing what it's called.

There are 7 Types of Nodes:

1. Root Node

   /

2. Element Node
3. Text Node text

   ()

4. Attribute Node

   @

5. Comment Node

   comment()

6. Processing Instruction

   processing-instruction()

7. Namespace Node

There are 13 AXIS Specifiers from your current node:

1. ancestor::

2. ancestor-or-self::

3. attribute::(or @)

4. child::(or nothing)

5. descendant::

6. descendant-or-self::

7. following::

8. following-sibling::

9. namespace::

10. parent::or (or ..)

11. preceding::

12. preceding-sibling::

13. self:: or (.)

Note that the double colon "::" immediately follows the use of an axis name.

Predicates:

Predicates are used to qualify expressions (i.e filter nodes). Predicates are specifiable on any node set and appear in sqare brackets after the node test.

Example:

version[3]

Functions:

• last()

  - When processing a list of nodes, if the nodes are numbered from one, last() returns the last number.

• position()

  - returns the value of the context position

• name()

  - returns the name of a node. Good for getting things like the element name for tests.

• concat() 

  - take two or more arguments and joins the results together end-to-end.

• count()

  - returns the number of nodes present in the node-set

• not()

  - returns the negation of its argument.

Examples:

Actually reading the XPath Expressions is probably the easiest way to learn them. Keep in mind that from any position in the document you can write an XPath expression to get to any other node in that document.

para

para/emphasis

//title

.//title 

item[1]

item[position()=last()]

topic[@level='basic']

//chapter[2]/title

preceding::partNumber

I get this request all the time so here is an example. It also shows you how to handle cases where you want to do something specific to certain elements while doing the copy.

<?xml version="1.0" encoding="UTF-8" ?>

<!--
    Document   : Create Copy of Document
    Created on : April 3, 2006
    Author     : sgmlxmldotnet
    Comment
       Demo file for Creating an Exact Copy of the document.
-->

<xsl:stylesheet version="1.0"
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
    <!-- template rule matching source root element -->
    <xsl:template match="/|*|comment()">
       <xsl:call-template name="do_copy"/>
    </xsl:template>

<!-- Add additional template matches here -->

<!-- Named Templates -->
	<!-- Copy current element and attribute nodes -->
	<xsl:template name="do_copy">
	    <xsl:copy>
        	<xsl:for-each select="@*">
				<xsl:copy/>
			</xsl:for-each>
			<xsl:apply-templates/>
        </xsl:copy>
    </xsl:template>

</xsl:stylesheet>

When your XSLT stylesheet seems to be doing strange things remember to
check the root element for a namespace. I banged my head against the
wall tonight for a good hour trying to figure out why I was getting
such strange results from simple template matches. If I did a match on
a wildcard like * it worked fine but if I did a specific match or xpath
expression nothing was coming back!

Example:

<myroot element xmlns="http://bogus.com/xmlbeans">

<some_elements/>

</myroot> 

Once you remember to check the root for a namespace life is fairly
simple just declare the namespace:

Stylesheet:

<xsl:stylesheet

                        xmlns:bogus="http://bogus.com/xmlbeans"

                        xmlns="http://bogus.com/xmlbeans">

</xsl:stylsheet>

Note that I declare the namespace twice. Once so I can address
it in the input file as "bogus:<some_element>" and then as the
default namespace for the resulting file so I don't have namespaces
declared all over the place.

Was recently working on an XSLT filter that extracted the text portion of SVG graphics into another file for translation. Once translated I needed a mechanism to merge the text back into the graphic. The document function did nicely.

Example:

<xsl:key name="descriptionlookup" 
match="Descriptions/Description" use="@name"/>
<xsl:variable name="desc" select="document('desc.xml')"/>

Later....

<xsl:template match="InsertDescription">
	<xsl:variable name="key_value" select="@lookup"/>
	<xsl:for-each select="$desc">
	<xsl:copy-of 
        select="key('descriptionlookup',$key_value)/Para"/>
	</xsl:for-each>
</xsl:template>

I was working on a stylesheet today that begs for a regular expression.

The Problem:

    <file path='connector\fig\82675-48040.svgz'/>

I want to use parts of the path attribute but since the length varies I
can't use the string() function. It appears there is no good way to do
it in XSLT 1.0

The XSLT 2.0 spec is a working draft at this point. It appears that in
2.0 something like this will be available:

    <xsl:value-of select='matches([REG EXP])'/>


CA



 This technique can be used to add metadata to incoming source documents.< ?xml:namespace prefix = o />

 Source:

 Stylesheet: