What Is XPath? Querying XML and HTML Documents

Short answer

XPath (XML Path Language) is a query language for selecting nodes in XML or HTML documents. It's like CSS selectors but more expressive — you can navigate parents, siblings, axes, and apply complex predicates. Browsers, scrapers, and test frameworks all support XPath natively.

Basic syntax

Expression	Selects
/html/body/div	Direct child path from root
//div	Any <div> in the document
//a[@href]	Any link with an href attribute
//a[@class="btn"]	Links with class exactly "btn"
//a[contains(@class,"btn")]	Links whose class contains "btn"
//ul/li[1]	First list item under any <ul>
//ul/li[last()]	Last list item
//div[text()="Hello"]	Div containing exact text "Hello"
//*[@id="main"]	Any element with id="main"
//input[@type="email"]	Email-type inputs
//button[contains(., "Submit")]	Button containing "Submit" anywhere in its text

Why XPath beats CSS selectors sometimes

	CSS Selectors	XPath
Match by text	No	Yes (text(), contains())
Walk to parent	No (without :has)	Yes (.., parent::)
Sibling navigation	Limited	Full (preceding/following axes)
Index from end	:nth-last-child	last(), last()-1
Predicates with logic	Comma OR only	and, or, not()
Speed in browsers	Faster	Slower
Readability	Cleaner	Verbose

Where XPath shines

• Test automation: Selenium and Playwright accept XPath; selectors based on visible text are dramatically more stable than CSS class hashes
• Web scraping: "find the next sibling of this label" is one XPath expression vs many lines of JS DOM walking
• XML processing: SOAP envelopes, RSS feeds, sitemaps, configuration files — all XPath-native
• Browser DevTools: in Chrome/Firefox console, $x("//a") evaluates an XPath; very useful for debugging

Common test selectors

// Click button labeled "Save"
//button[normalize-space()="Save"]

// Find input following a label "Email"
//label[text()="Email"]/following-sibling::input[1]

// Find row in a table by cell content
//tr[td[contains(., "Order #1234")]]

// Cell to the right of "Total:" label
//td[text()="Total:"]/following-sibling::td[1]

Versions

• XPath 1.0 (1999) — what browsers support, simple syntax
• XPath 2.0/3.0/3.1 — added regex, sequences, math; only XSLT engines (Saxon) support these

For browser/Selenium/Playwright work, you're using XPath 1.0.

JSON equivalent

XPath is XML-specific. For querying JSON, use JSONPath — same idea, JSON syntax. Try our JSONPath tester.

Related tools

Test patterns against text (XPath has limited regex; sometimes you need real regex): regex tester. Convert HTML entities encountered in scraped text: HTML encoder/decoder.