XSL using index of node at any level

Question

I have the following xml:

<content> <p>para 1<an>test 1</an></p> <p>para 2<an>test 2</an></p> <p>para 3<an>test 3</an></p> </content>

and I have the following expression in xsl:

<xsl:template match="/"> <xsl:text>Count: </xsl:text> <xsl:value-of select="count(/content//an)" /> <xsl:text> Content: </xsl:text> <xsl:value-of select="/content//an[2]" /> </xsl:template>

Why does /content//an[1] returns "test 1" and /content//an[2] return ""? The count says that there are 3. I'm using libxslt.

Thanks!

Good question, +1. See my answer for an explanation and complete and easy solution. — Dimitre Novatchev
– Dimitre Novatchev, Commented Mar 22, 2011 at 13:33
possible duplicate of Why does //span[2] not select the second span in a document? — user357812
– user357812, Commented Mar 22, 2011 at 22:57

Dimitre Novatchev · Accepted Answer · 2011-03-22 14:34:58Z

This is a FAQ.

One must be careful when using the // abbreviation.

//SomeName[1]

means: select all SomeName elements in the document that are the first SomeName children of their parent. Often the selected nodes are more than one (or even all the nodes).

The correct way to select only the $k-th SomeName element in the document is:

(//SomeName)[$k]

In your case, use:

(/content//an)[2]

Remember: The [] operator has a higher precedence (binds stronger) than the // abbreviation.

Collectives™ on Stack Overflow

XSL using index of node at any level

1 Answer 1

1 Comment

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Related