What Is xsl:analyze-string in XSLT 2.0?

The <xsl:analyze-string> instruction in XSLT 2.0 provides native regular expression processing to parse, tokenize, and transform unstructured text into structured XML markup. Prior to XSLT 2.0, string parsing required verbose, recursive named templates using basic functions like substring-before and substring-after. With <xsl:analyze-string>, developers can evaluate regular expressions against any string expression and cleanly separate matching segments from non-matching segments to apply targeted transformations.

Core Structure and Syntax

The instruction operates on an input string defined by the select attribute and applies a standard regular expression defined by the regex attribute. It contains two optional child elements that determine how matched and unmatched substrings are processed:

<xsl:matching-substring>: Executes its sequence constructor for every portion of the input string that matches the regular expression.
<xsl:non-matching-substring>: Executes its sequence constructor for text segments between matches.

<xsl:analyze-string select="description" regex="[A-Z]{3}-\d{4}">
  <xsl:matching-substring>
    <product-code><xsl:value-of select="."/></product-code>
  </xsl:matching-substring>
  <xsl:non-matching-substring>
    <xsl:value-of select="."/>
  </xsl:non-matching-substring>
</xsl:analyze-string>

Within both child blocks, the context item (.) evaluates to the string representing the current matched or non-matched fragment.

Accessing Capturing Groups with regex-group()

A major advantage of <xsl:analyze-string> over simple regex boolean functions is the ability to extract sub-matches using parentheses in the pattern. Inside <xsl:matching-substring>, the regex-group($n) function retrieves the substring captured by the $n$-th parenthesized group.

For example, when parsing dates formatted as YYYY-MM-DD:

<xsl:analyze-string select="raw-date" regex="^(\d{4})-(\d{2})-(\d{2})$">
  <xsl:matching-substring>
    <date>
      <year><xsl:value-of select="regex-group(1)"/></year>
      <month><xsl:value-of select="regex-group(2)"/></month>
      <day><xsl:value-of select="regex-group(3)"/></day>
    </date>
  </xsl:matching-substring>
</xsl:analyze-string>

Calling regex-group(0) returns the entire matched substring, identical to evaluating the context item ..

Regex Flags and Modifiers

The optional flags attribute allows configuration of the regex evaluation engine using single-character flags:

s (dot-all mode): Allows the . wildcard to match any character, including newline characters (\n and \r).
m (multi-line mode): Treats ^ and $ as anchors for the start and end of individual lines rather than the entire input string.
i (case-insensitive mode): Matches alphabetic characters regardless of case.
x (extended mode): Ignores whitespace within the regex string, allowing patterns to be formatted with comments for readability.

Practical Applications in XML Pipelines

Converting Plain Text to Semantic Markup

When importing legacy documents or plain text blocks, <xsl:analyze-string> can identify URLs, email addresses, or specific code identifiers and wrap them in proper XML tags without affecting surrounding prose.

Splitting Delimited Content

While tokenize() is suitable for splitting strings into atomic items, <xsl:analyze-string> is essential when the delimiter itself contains metadata or when variable delimiters require distinct markup wrappers.

Normalizing Mixed-Format Data

In data migration scenarios where data fields contain mixed alphanumeric codes, regular expressions can separate prefix codes, sequence numbers, and suffixes into discrete XML attributes or child nodes in a single pass.