Re: Ignoring quotation marks in grouping and sorting

Hi Syd --

I tried to post my XML and XSL to the list but it got rejected for size.  Will
send to you offlist...

Michele

-----Original Message-----
From: Syd Bauman s.bauman@xxxxxxxxxxxxxxxx
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Sent: Monday, January 7, 2019 10:23 AM
To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
Subject: Re:  Ignoring quotation marks in grouping and sorting

Oooh. I had completely forgotten that FMP has its own internal XSLT transform
capability. Yes, to my (limited) knowledge, Xalan is an XSLT
1.0 processor.

As for the "getValue" template I showed you (off-list for those who are
confused), it is designed to be called once *from the template processing an
individual FMPXMLRESULT row*, and used to set a variable when called. Then any
subsequent references are via to that variable.

But if what you are doing is sorting rows, as it were, that may not work for
you. I am not sure w/o looking at your program (and even then may not be sure.
:-) if there is an easy way around that. (IIRC, <xsl:key>, at least in XSLT
1.0, cannot have children. I.e., you cannot use construction elements as the
content of <xsl:key> in place of @use.)

Would be very helpful if you could post your complete input (perhaps with only
a few lines of fake data if you don't want to reveal your real data to the
world), and the corresponding desired output. You may want to post your XSLT
1.0 you have so far, as well.

All that said, access by field name (which is what "getValue" does) is easier
to work with and more robust, but if using COL[2] and an <xsl:sort> is working
for you, it may be just fine to leave well enough alone. :-)

> Thanks, Syd, the "getValue" template looks very useful. Am I right in
> thinking that I would need to call this template all three places
> -- i.e., where I'm setting up the keys, where I do the grouping for
> output, and where I sort the output for each letter? Because the
> leading quote needs to be ignored in all three places.
>
> Also, just FYI, I was wrong about the XSLT processor. We use Saxon for
> our standalone transforms but this one happens as part of a FileMaker
> export (FMP offers the option of applying a style sheet as part of the
> XML export process). According to the documentation, "...The XML
> import and export features in FileMaker Pro use an XML parser called
> Xerces and an XSLT style sheet processor called Xalan." From what I
> can glean from the FileMaker community, it's rather old tech and
> doesn't support 2.0 let alone 3.0