lang()

Description

lang() tests whether the context node is written in the language identified by the argument string. It walks up the ancestor-or-self axis to find the nearest xml:lang attribute and then compares its value to the argument using a case-insensitive prefix match.

The matching rule is: if the xml:lang value equals the argument (case-insensitive), or if it equals the argument followed by a hyphen (-) and any subtag, then lang() returns true. For example, lang('en') returns true for nodes where xml:lang is en, en-US, en-GB, or EN-AU.

This makes lang() well-suited for filtering multilingual documents by language family without needing to enumerate every regional variant explicitly.

Parameters

Return value

xs:boolean — true if the context node’s effective xml:lang matches the argument, false otherwise.

Examples

Filter paragraphs by language

Input XML:

<?xml version="1.0" encoding="UTF-8"?>
<doc>
  <para xml:lang="en">Hello, world.</para>
  <para xml:lang="fr">Bonjour le monde.</para>
  <para xml:lang="en-GB">Cheers, mate.</para>
</doc>

Stylesheet:

<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <xsl:output method="xml" indent="yes"/>

  <xsl:template match="/doc">
    <english>
      <xsl:for-each select="para[lang('en')]">
        <p><xsl:value-of select="."/></p>
      </xsl:for-each>
    </english>
  </xsl:template>
</xsl:stylesheet>

Output:

<english>
  <p>Hello, world.</p>
  <p>Cheers, mate.</p>
</english>

Inherited xml:lang from ancestor

Input XML:

<?xml version="1.0" encoding="UTF-8"?>
<book xml:lang="de">
  <chapter>
    <title>Einleitung</title>
    <para>Ein einleitender Absatz.</para>
    <para xml:lang="en">An English aside.</para>
  </chapter>
</book>

Stylesheet:

<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <xsl:output method="xml" indent="yes"/>

  <xsl:template match="/book">
    <german-content>
      <xsl:for-each select="//para[lang('de')]">
        <item><xsl:value-of select="."/></item>
      </xsl:for-each>
    </german-content>
  </xsl:template>
</xsl:stylesheet>

Output:

<german-content>
  <item>Ein einleitender Absatz.</item>
</german-content>

Notes

• The comparison is case-insensitive: lang('EN') and lang('en') behave identically.
• lang() searches the nearest ancestor-or-self that carries an xml:lang attribute. If no such ancestor exists, the function returns false.
• The argument is matched as a prefix: lang('zh') matches zh-Hant and zh-Hans but not zho.
• lang() only recognises the xml:lang attribute in the XML namespace. A plain lang attribute without the xml: prefix is ignored.
• In XSLT 2.0+, lang() is still available with the same semantics; additionally, the xsl:sort element’s lang attribute drives language-sensitive collation separately.

See also

• normalize-space()
• string()