Thus far (if you've been following the pre-defined primer structure), we have covered how to navigate the input document tree using simple XPath expressions to select particular elements and transform XML documents. Now we will see how to process and transform your input documents in more complex ways using conditional branching structures. Many of these examples will make use of XSL variables. This slide provides the syntax required to set and reference XSL variables within an XSL stylesheet.

If you download and open the two pages in Oxygen (or open the examples in a browser and view the page source), you can see some of the benefits of using XSL variables.

While both stylesheets are functionally the same, we can see the ease with which the number of authors, rulers, and people can be called upon and manipulated when using XSL variables. You might have a much longer XSL document that calls upon the same XPath expressions over and over again. If this is the case, you may want to use XSL variables. Another benefit to using XSL variables is that they can help other people working on the project read and understand your XSL stylesheets (provided you name your XSL variables intuitively).

Generally, XSLT templates are representations of contingency; they essentially say "when this node is matched, do this." However, there are cases where these tests are not sufficient. Instead, we want to set up a set of conditions: When A do a; when B, do b, rather than simply matching one node to the other. It is usually inconvenient to do this simply through matching nodes, so we use conditionals to express what we want. Some examples of why you may want to use conditionals include, adding an s to something that is plural, adding a line number to every fifth line, or use a default for missing data.

XSLT has two conditional branching structures. These are xsl:if and xsl:choose. These two structures will be covered in this tutorial.

If you would like to learn about the XPath if-then-else structure (and there is no reason why you shouldn't!), you can refer to these slides.

The example in this slide uses xsl:if to pluralize the word "author" if there is more than one listed. The slide explains in detail how this particular xsl:if is evaluated.

In general, every test attribute will contain an XPath expression. The expression either evaluates as true or as false. If it is true, then the content of xsl:if is applied. If it is false, nothing is done.

In this example, the content of xsl:if is a simple text node, but we can imagine more complex examples like checking if there is an ref on persName, and if there is, then proceeding with a whole additional snippet of code.

The xsl:if element works well if you only want to apply constructors (i.e., do things) when a simple test proves to be true (or false, as you can always reverse the logic of a test by wrapping it in a not() function), and don't want to do anything if said simple test proves to be false (or true). But there may be instances in which you want to do something when the test evaluates as true, and something else when it evaluates as false. Indeed, you may want to apply rules to things that aren't even simple true/false statements.

In these cases, you will want to use xsl:choose, which allows you to specify rules for several different scenarios. The children of xsl:choose are one or more xsl:when statements followed by zero or one xsl:otherwise statement.

The processor evaluates the test attribute of each xsl:when in order. If the XPath expression is false, the processor goes to the next xsl:when. When the processor finds that the test of an xsl:when is true, it performs the content of that xsl:when, and stops looking at the rest of the xsl:when elements and the xsl:otherwise (if there even is one).

If none of the test attributes for xsl:when are true, then the processor will execute the content of the xsl:otherwise (if there is one).

Thus at most one child of an xsl:choose is executed. If there is no xsl:otherwise child of the xsl:choose, it is possible that no child of the xsl:choose will be executed.

Most situations you encounter will have more than one way to process your documents in the ways that you want. Here, each example accomplishes the same thing. Whether you use one or the other is a matter of preference or convenience. It is possible that one is faster than the other using a particular XSLT engine; but the other might be faster than the one using a different XSLT engine. And the difference is probably infinitesmal, anyway.

The XPath if-then-else tutorial includes some exercise files with which you can practice using conditional brancing structures. It is also the next tutorial in the primer sequence.