saxon:analyze-string

The action of this function is analagous to the xsl:analyze-string instruction in XSLT 2.0. It is provided to give XQuery users access to regular expression facilities comparable to those provided in XSLT 2.0. (The function is available in XSLT also, but is unnecessary in that environment.)

The first argument defines the string to be analyzed. The second argument is the regex itself, supplied in the form of a string: it must conform to the same syntax as that defined for the standard XPath 2.0 functions such as matches().

The third and fourth arguments are function items, called the matching and non-matching functions respectively. The matching function is called once for each substring of the input string that matches the regular expression; the non-matching function is called once for each substring that does not match. These functions may return any sequence. The final result of the saxon:analyze-string function is the result of concatenating these sequences in order.

The matching function takes two arguments. The first argument is the substring that was matched. The second argument is a sequence, containing the matched subgroups within this substring. The first item in this sequence corresponds to the value $1 as supplied to the replace() function, the second item to $2, and so on.

The non-matching function takes a single argument, namely the substring that was not matched.

The detailed rules follow xsl:analyze-string. The regex must not match a zero-length string, and neither the matching nor non-matching functions will ever be called to process a zero-length string.

The following example is a "multiple match" example. It takes input like this:

<doc>There was a young fellow called Marlowe</doc>

and produces output like this:

<out>Th[e]r[e] was a young f[e]llow call[e]d Marlow[e]</out>

The XQuery code to achieve this is:

declare namespace f="f.uri"; declare function f:match ($c, $gps) { concat("[", $c, "]") }; declare function f:non-match ($c) { $c }; <out> { string-join( saxon:analyze-string(doc, "e", f:match#2, f:non-match#1)), "") } </out>

The following example is a "single match" example. Here the regex matches the entire input, and the matching function uses the subgroups to rearrange the result. The input in this case is the document <doc>12 April 2004</doc> and the output is <doc>2004 April 12</doc>. Here is the query:

declare namespace f="f.uri"; declare function f:match ($c, $gps) { string-join(($gps[3], $gps[2], $gps[1]), " ") }; declare function f:non-match ($c) { error("invalid date") }; <out> { saxon:analyze-string(doc, "([0-9][0-9]) ([A-Z]*) ([0-9]{4})", f:match#2, f:non-match#1, "i") } </out>

This particular example could be achieved using the replace() function: the difference is that saxon:analyze-string can insert markup into the result, which replace() cannot do.