Test only CDATA content string with XPath

Question

I have the following XML code :

<?xml version="1.0" encoding="ISO-8859-1"?>
<articles>
    <article>
        <title>
            <![CDATA[The title of the test]]>
        </title>
        <date>
            <![CDATA[27/07/2013]]>
        </date>
        <url>
            <![CDATA[http://www.google.com]]>
        </url>
    </article>
    <article>
        <title>
            <![CDATA[The title of the SECOND test]]>
        </title>
        <date>
            <![CDATA[27/12/2013]]>
        </date>
        <url>
            <![CDATA[]]>
        </url>
    </article>
</articles>

I want to test if the string inside URL's CDATA is empty. In my XML example, I consider that the first URL node isn't empty but the second is an empty string.

Is it possible to test this with XPath in an XSL 1.0 stylesheet (I need to parse this XML with XSL) ?

I tried with the following XPath selectors :

string-length(/articles/article[2]/url)
string-length(/articles/article[2]/url/text())

The first selector result is 7, according to XML Tools in Notepad++, and 4 for the second one.

Answer 1

Use the normalize-space function to strip starting and trailing whitespaces: http://www.w3.org/TR/xpath/#function-normalize-space

like this

string-length(normalize-space(/articles/article[2]/url))