[XSL-LIST Mailing List Archive Home]
[By Thread]
[By Date]
RE: [xsl] XSLT 2.0 function - fastest node comparison
Subject: RE: [xsl] XSLT 2.0 function - fastest node comparison From: "Michael Kay" <mike@xxxxxxxxxxxx> Date: Thu, 10 Mar 2005 16:21:12 -0000 |
> Interestingly, using 'cast as' on $middle to narrow it to an integer > causes the execution time to increase by 250ms, so I've left > them all as > doubles (the positional predicate in $currentNode doesn't > appear to mind > using a double). That makes me realise that it would actually avoid many string->number conversions if you extracted the ranges as two sequences of integers in global variables: <xsl:variable name="lows" select="$ranges/*/@from/xs:integer(.)" as="xs:integer*"/> <xsl:variable name="highs" select="$ranges/*/@to/xs:integer(.)" as="xs:integer*"/> <xsl:variable name="n" select="count($lows)" as="xs:integer"/> test="some $i in 1 to $n satisfies ($highs[$i] ge $test and $lows[$i] le $test)" Michael Kay http://www.saxonica.com/
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
RE: [xsl] XSLT 2.0 function - faste, Andrew Welch | Thread | RE: [xsl] XSLT 2.0 function - faste, Andrew Welch |
Re: [xsl] RTF from an html, Stef | Date | Re: [xsl] calling two consecutive x, RQamar |
Month |