XPath Tutorial — Query XML Documents with Path Expressions

XPath (XML Path Language) is a query language for selecting nodes from XML documents. It uses path-like syntax to navigate the element tree, similar to file system paths or CSS selectors — but with more expressive filtering via predicates and functions.

Use the XML Formatter to inspect and work with XML structures.

XPath fundamentals

Given this XML:

<bookstore>
  <book category="fiction">
    <title lang="en">The Great Gatsby</title>
    <author>F. Scott Fitzgerald</author>
    <price>12.99</price>
  </book>
  <book category="tech">
    <title lang="en">Clean Code</title>
    <author>Robert C. Martin</author>
    <price>34.99</price>
  </book>
  <book category="fiction">
    <title lang="fr">Le Petit Prince</title>
    <author>Antoine de Saint-Exupéry</author>
    <price>9.99</price>
  </book>
</bookstore>

Basic XPath expressions:

Expression	Selects
/bookstore	Root element bookstore
/bookstore/book	All book elements that are direct children of bookstore
//book	All book elements anywhere in the document
//title	All title elements
//book/title	All title elements that are children of book
//@lang	All lang attributes
//book[1]	First book element
//book[last()]	Last book element

Predicates (filtering)

Predicates in square brackets filter nodes:

//book[@category='fiction']

→ All book elements where category attribute equals “fiction”

//book[price > 20]

→ All books with price greater than 20

//book[author='Robert C. Martin']

→ Books by Robert C. Martin

//book[@category='tech' and price < 40]

→ Tech books under $40

//title[@lang='en']

→ Titles in English

//book[position() <= 2]

→ First two books

Axes

XPath axes select nodes relative to the context node:

Axis	Meaning
child::	Direct children (default)
parent::	Parent element
ancestor::	All ancestors (parent, grandparent, …)
descendant::	All descendants
following-sibling::	Siblings after context node
preceding-sibling::	Siblings before context node
attribute::	Attributes of context node
self::	Context node itself

//book/following-sibling::book

→ All book elements that come after each book element

//title/parent::book

→ book elements that contain title elements

//book/ancestor::bookstore

→ The bookstore element (ancestor of any book)

Abbreviated syntax:

• . = self::node()
• .. = parent::node()
• @attr = attribute::attr
• // = /descendant-or-self::node()/

Functions

String functions

//title[contains(text(), 'Clean')]

→ Titles containing “Clean”

//title[starts-with(text(), 'The')]

→ Titles starting with “The”

//book[string-length(title) > 10]

→ Books with titles longer than 10 characters

normalize-space(//title[1])

→ First title with leading/trailing whitespace removed

Numeric functions

count(//book)

→ Number of book elements

sum(//price)

→ Sum of all prices

//book[price = min(//price)]

→ Cheapest book (XPath 2.0)

Node functions

//book[name()='book']

→ Elements named “book” (redundant here but useful dynamically)

//book[not(@category)]

→ Books without a category attribute

//book[@category != 'fiction']

→ Books where category is not fiction (includes books without category)

XPath in JavaScript (browser)

const xml = `<bookstore>
  <book category="fiction"><title>Gatsby</title><price>12.99</price></book>
  <book category="tech"><title>Clean Code</title><price>34.99</price></book>
</bookstore>`;

const parser = new DOMParser();
const doc = parser.parseFromString(xml, 'application/xml');

// Evaluate XPath:
function xpathQuery(expression, doc) {
  const result = doc.evaluate(
    expression,
    doc,
    null,  // Namespace resolver
    XPathResult.ANY_TYPE,
    null
  );
  
  const nodes = [];
  let node = result.iterateNext();
  while (node) {
    nodes.push(node);
    node = result.iterateNext();
  }
  return nodes;
}

// Get all book titles:
const titles = xpathQuery('//title/text()', doc);
titles.forEach(t => console.log(t.textContent));
// "Gatsby"
// "Clean Code"

// Get fiction books:
const fiction = xpathQuery('//book[@category="fiction"]', doc);
console.log(fiction.length);  // 1

// Get price of tech books:
const techPrices = xpathQuery('//book[@category="tech"]/price/text()', doc);
console.log(techPrices[0].textContent);  // "34.99"

XPath in Python (lxml)

from lxml import etree

xml = b'''<bookstore>
  <book category="fiction"><title>Gatsby</title><price>12.99</price></book>
  <book category="tech"><title>Clean Code</title><price>34.99</price></book>
</bookstore>'''

root = etree.fromstring(xml)

# Simple selection:
titles = root.xpath('//title/text()')
print(titles)  # ['Gatsby', 'Clean Code']

# Filter by attribute:
fiction_books = root.xpath('//book[@category="fiction"]')
for book in fiction_books:
    print(book.find('title').text)  # "Gatsby"

# Numeric result:
total_price = root.xpath('sum(//price)')
print(total_price)  # 46.98

count = root.xpath('count(//book)')
print(int(count))  # 2

# With namespaces:
xml_ns = b'''<root xmlns:dc="http://purl.org/dc/elements/1.1/">
  <dc:title>My Doc</dc:title>
</root>'''
root_ns = etree.fromstring(xml_ns)
titles = root_ns.xpath('//dc:title/text()', 
                       namespaces={'dc': 'http://purl.org/dc/elements/1.1/'})
print(titles)  # ['My Doc']

XPath vs CSS selectors

Task	XPath	CSS Selector
All descendants	//div	div
Direct child	/div/p	div > p
By attribute	//a[@href]	a[href]
By attribute value	//a[@class='nav']	a.nav or a[class='nav']
First element	(//li)[1]	li:first-child
Last element	(//li)[last()]	li:last-child
Contains text	//p[contains(text(),'hello')]	No equivalent
Parent selection	//li/..	No equivalent
Sibling selection	//h2/following-sibling::p[1]	h2 + p
Select by text	//button[text()='Submit']	No equivalent

XPath is more powerful for data extraction (can select parent nodes, filter by text content). CSS selectors are more concise for styling and simpler DOM traversal.

Related tools

• XML Formatter — format and inspect XML documents
• XML Validator Online — validate XML syntax
• XML Namespace Guide — using XML namespaces