RE: [xsl] select first word

Subject: RE: [xsl] select first word
From: "Scott Trenda" <Scott.Trenda@xxxxxxxx>
Date: Tue, 18 Dec 2007 11:33:00 -0600
If we don't care about preserving the commas, we could just translate them to
spaces to start with, and simply check for the space after that. Then it's not
dependent on a space being followed by a comma. Something like this:

<xsl:variable name="trans" select="translate(.,
'abcdehijklmnorstuvyz{|HIJKA@,', 'aaaaaeeeeiiiioooooEEEEAA ')" />
	<xsl:when test="contains($trans, ' ')">
		<xsl:value-of select="substring-before($trans, ' ')" />
		<xsl:value-of select="$trans" />

~ Scott

-----Original Message-----
From: Abel Braaksma []
Sent: Tuesday, December 18, 2007 11:12 AM
To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
Subject: Re: [xsl] select first word

Florent Georges wrote:
> Abel Braaksma wrote:
>> <xsl:value-of select="
>>         translate(
>>            substring-before(
>>                substring-before(.,' '), ', '),
>>         $from, $to)" />
>   If I am right (if I remember correctly) the input doesn't always have
> space or comma.  So I would use the following instead (or the result
> could be empty):
>     translate(replace(., '([^, ]+).*', '$1'), $from, $to)

Thanks for pointing that out.
Yes, you are right about the input. But the OP has, I believe (from
observation), only XSLT 1.0 and replace() is introduced as per XSLT 2.0.

The substring-before with the comma can be removed: if there's a comma
it is followed by a space. Just add the comma to the translate function
at the end, then it will be removed.

For the rest I'd probably choose something like this:

<!-- include comma, which is not matched, thus it will be deleted -->
<xsl:variable name="from" select="'abcdehijklmnorstuvyz{|HIJKA@,'" />
<xsl:variable name="to" select="'aaaaaeeeeiiiioooooEEEEAA'" />

<!-- with space -->
<xsl:template match="le[contains(., ' ')]">
    <xsl:value-of select="translate(substring-before(., ' '), $from,
$to)" />

<!-- without space -->
<xsl:template match="le">
    <xsl:value-of select="translate(., $from, $to)" />

The second one is the more generic one and will be called by the
processor if there's no space in the content of the <le> element.


-- Abel Braaksma

Current Thread