XPath find if node exists

Question

Using a XPath query how do you find if a node (tag) exists at all?

For example if I needed to make sure a website page has the correct basic structure like /html/body and /html/head/title.

Libre

Maybe it's better to use XML Schema with obligatory elements indication? So check that a document uses it or not. – Cockneyism

Answer 1

<xsl:if test="xpath-expression">...</xsl:if>

so for example

<xsl:if test="/html/body">body node exists</xsl:if>
<xsl:if test="not(/html/body)">body node missing</xsl:if>

Furnishings

@SearchForKnowledge, you should probably ask that as a new question at SO, but as a quick guide: html/body and not(html/body/node()) (i.e., just test if it exists and it does not contain any child nodes, or text nodes). – Betook

Answer 2

Try the following expression: boolean(path-to-node)

Bedrock

This is exactly what I needed when using XPath in Python with lxml. – Franklin

This worked a treat when constructing Rule Conditions and Rule Actions within InfoPath 2010. – Mucilage

This is the correct answer in the sense that this is actually an XPath query, unlike most of the other answers. – Festivity

Answer 3

Patrick is correct, both in the use of the xsl:if, and in the syntax for checking for the existence of a node. However, as Patrick's response implies, there is no xsl equivalent to if-then-else, so if you are looking for something more like an if-then-else, you're normally better off using xsl:choose and xsl:otherwise. So, Patrick's example syntax will work, but this is an alternative:

<xsl:choose>
 <xsl:when test="/html/body">body node exists</xsl:when>
 <xsl:otherwise>body node missing</xsl:otherwise>
</xsl:choose>

Rapture

+1 for noting if-then-else, also what about if-else if-else? In davenpcj's answer can I place test="somexpath" in the 2nd when to make it if-else if-else? – Guthrie

@Abhishek Yes, you can put more xsl:when with other conditions and have a multi-branch statement. Think of it more like a SELECT than an if-then-else, with xsl:otherwise as the default:. – Striker

This is awesome but what if I wanted to check if it exists OR empty? – Billbillabong

Answer 4

Might be better to use a choice, don't have to type (or possibly mistype) your expressions more than once, and allows you to follow additional different behaviors.

I very often use count(/html/body) = 0, as the specific number of nodes is more interesting than the set. For example... when there is unexpectedly more than 1 node that matches your expression.

<xsl:choose>
    <xsl:when test="/html/body">
         <!-- Found the node(s) -->
    </xsl:when>
    <!-- more xsl:when here, if needed -->
    <xsl:otherwise>
         <!-- No node exists -->
    </xsl:otherwise>
</xsl:choose>

Striker

As indicated in the code above, more xsl:when clauses can be added to change the behavior and handle multiple conditions in different ways. – Striker

count(/html/body) = 0 genius ! :D I use it as /html[count(/body)=0]/someNode for selecting someNode when /body (or whatever) is missing – Stakeout

@clickstefan, /html[count(/body)=0] will never select anything, there cannot be two root nodes in XML. Maybe you meant /html[count(body)=0], which would be the same as /html[not(body)], or /html[not(exists(body))]. – Betook

@Betook yes, /html[count(//body)=0] or how you said, shameful but yes I didn't give a proper example – Stakeout

Answer 5

I work in Ruby and using Nokogiri I fetch the element and look to see if the result is nil.

require 'nokogiri'

url = "http://somthing.com/resource"

resp = Nokogiri::XML(open(url))

first_name = resp.xpath("/movies/actors/actor[1]/first-name")

puts "first-name not found" if first_name.nil?

Barbabas

Answer 6

A variation when using xpath in Java using count():

int numberofbodies = Integer.parseInt((String) xPath.evaluate("count(/html/body)", doc));
if( numberofbodies==0) {
    // body node missing
}

Renatarenate