[xsl] Re: Getting the offset of a matching-substring in <xsl:analyze-string>

[Martin Holmes <mholmes@xxxxxxx>]

On 12-03-16 02:09 PM, Michael Kay wrote:
> The new fn:analyze-string() function in XPath 3.0, designed to deliver
> the benefits of xsl:analyze-string to XQuery users, produces as output a
> copy of the input string with added markup to identify the matching and
> non-matching substrings, and the captured groups. So you can examine
> this output to see where the captured groups are in the string; but it's
> still a second pass of a kind.
>
> http://www.w3.org/TR/xpath-functions-30/#func-analyze-string

It's pretty much _exactly_ what I want -- in fact, I ended up building a 
structure that's almost identical to analyze-string-result but with 
different tag names.

Off to read the rest of the spec...

Cheers,
Martin

> Michael Kay
> Saxonica
>
> On 16/03/2012 20:57, Martin Holmes wrote:
> > That's what I did in the end. It would be nice not to have to, though.
> >
> > Cheers,
> > Martin
> >
> > On 12-03-15 05:49 PM, Imsieke, Gerrit, le-tex wrote:
> > > I think you need a second pass to calculate the offsets.
> > >
> > > On 2012-03-16 01:00, Martin Holmes wrote:
> > > > Hi there,
> > > >
> > > > I'm doing some automated markup using <xsl:analyze-string>, and as part
> > > > of the output I'd like to record the offset and length of each matched
> > > > substring. So given input like this in a text node (simplified
> > > > example):
> > > >
> > > > This is a test.
> > > >
> > > > and a match like this:
> > > >
> > > > <xsl:analyze-string select="$input" regex="{'[a-zA-Z]+'}">
> > > >
> > > > <xsl:matching-substring>
> > > >
> > > > <w offset="???" length="{string-length(.)}">
> > > >
> > > > <xsl:value-of select="."/>
> > > >
> > > > </w>
> > > >
> > > > <xsl>matching-substring>
> > > >
> > > > </xsl:analyze-string>
> > > >
> > > > I'm wondering if there's any way I can calculate the offset value
> > > > (which
> > > > for "This" would be 1, for "is" would be 6, etc.).
> > > >
> > > > Does anyone know if there's a way to do this?
> > > >
> > > > All help appreciated,
> > > > Martin