<xsl:element>

Up to now, when we have wanted to include an element in the result tree, we have just gone ahead and typed the start and end tags into the stylesheet. In most cases, that works. However, it will not work if the element name depends on data in the source XML document.

 

In this case we use <xsl:element> to create the element for us. This is used most frequently when creating an element-rich result tree from an attribute-rich source tree. However, we don't have an attribute-rich document to play with, so we will simply create element names based on the id attributes of the books in our catalog. We cannot use the id values themselves as element names. (Why not? Think back to your XML specification or wait to see at the end of this section.) So we will use the following code (id.xsl):

 

Notice the line:

 

 

The part in braces is an attribute value template, a feature of XSLT we will be meeting in the next chapter. This is the required format for the value of the name attribute.

 

Notice also the use of <xsl:text>:

 

 

The purpose of this was to put a line break between each element of the output. This is only there to make the resulting serialized document more readable. And here it is (when applied to catalog.xml):

 

 

There is another occasion you might want to use <xsl:element>. Because the element allows a namespace attribute, it provides more control over the namespace than we have when including the element as literal text in the stylesheet. Just as we have made the element name in the result tree dependent on the content of the input tree, we can do the same with a namespace. Like the name attribute, the value of this attribute is an attribute value template.

 

The final, optional, attribute of <xsl:element>, called use-attribute-sets, allows us to specify an attribute set to use with the defined element. Such sets of attributes can be defined using the <xsl:attribute-set> element.

 

Why couldn't we use the id values themselves as element names earlier? Our id values start with a digit – remember from the XML specification that an element name must start with an alphabetic character or an underscore or colon. Of course, we would never use a colon as this is used as a separator between a namespace prefix and local name.

<xsl:attribute>

You will probably find that you use <xsl:attribute> more often than <xsl:element>, particularly if you are creating HTML output.

 

This element allows us to create an attribute name and/or value in the result tree based on data in the source tree. For example, in our book catalog we have references to images of the book covers, such as:

 

 

If we were to display information from the catalog by creating an HTML page, we would want to create an <IMG> element referring to this image. Something like:

 

 

We can do this using an attribute value template. The following stylesheet, covers.xsl, creates a table of book titles and images:

 

 

Note that we have used the line:

 

This ensures that the resulting <IMG> element does not use the empty element syntax. The serialized output will have an element of the form:

 

 

rather than:

 

 

This ensures compatibility with a greater range of older browsers.

 

The part we are most interested in is:

 

 

Note that normally I would put the <xsl:attribute> element and all its content on a single line. This is because this element contains a text node (covers/), and so the additional whitespace will be included in the output. I have only reformatted this line to make it fit within the constraints of the printed page.

 

All the code above is doing is creating the <IMG> element in the output with a src attribute using the value of the text node of the <CoverImage> element in the source tree. In this case, we have included some additional text to provide the correct relative URL for the images.

 

This screenshot shows the first four books:

 

 

Note that you would not use both an <xsl:attribute> and an <xsl:element> element. Instead you would use an attribute value template.

 

We have now covered the main elements of XSLT. With these, you can write many useful stylesheets. We will next look at two important aspects of XSLT – default templates and what happens when several templates match the same pattern.