Re: First Element in Sorted List
On 6/2/05, Michael Kay <mike@xxxxxxxxxxxx> wrote: > > > > My process needs to take action on the first element in a list of > > elements where the logic to determine first is the lowest > > numeric value > > of the sortorder attribute. I don't believe there is a > > one-line way to > > do this - I believe the entire list must be sorted, and then the > > position of each element after sort compared to determine if > > it is first. > > Yes. Even in XSLT 2.0 this may be the best approach. 2.0 offers min/max > functions, but they give you the highest/lowest value, not the node having > that value. > > Writing a recursive template/function might give you better performance than > sorting (linear rather than n*log(n)). > > There is a one-line solution: $x[not($x/@sortorder < @sortorder)] but it's > quite likely to have poor (quadratic) performance. > > You could consider using EXSLT math:highest() or math:lowest(). > > Or I'm sure there's something in Dimitre's box of tricks called FXSL. > > Michael Kay > http://www.saxonica.com/ The following FXSL one-liner does the job: <xsl:sequence select= "2*f:xsltSort(/*/*/@sortorder/xs:integer(.), f:add(0))"/> In this case, the "do-something" operation on the minimum item of the sequence is double-it. When evaluated against the following source xml: <t> <x sortorder="3"/> <x sortorder="2"/> <x sortorder="5"/> <x sortorder="8"/> <x sortorder="7"/> <x sortorder="9"/> <x sortorder="6"/> <x sortorder="4"/> <x sortorder="1"/> <x sortorder="2"/> </t> The correct result: 2 is produced. The f:xsltSort() function is simply a wrapper around the xsl:perform-sort instruction. It accepts a list of functions, which are composed to produce the effect of multiple sort keys. Here's the code of the function: <xsl:function name="f:xsltSort" as="item()*"> <xsl:param name="pSeq" as="item()*"/> <xsl:param name="pCriteria" as="node()*"/> <xsl:perform-sort select="$pSeq"> <xsl:sort select="f:compose-flist($pCriteria, .)"/> </xsl:perform-sort> </xsl:function> Certainly, in a future refinement I should also either process an explicitly passed type-argument (e.g. number or text) or find a way to efficiently determine the common most non-generic type of all items in a sequence. Any ideas? -- Cheers, Dimitre Novatchev --------------------------------------- To avoid situations in which you might make mistakes may be the biggest mistake of all.
PURCHASE STYLUS STUDIO ONLINE TODAY!
Purchasing Stylus Studio from our online shop is Easy, Secure and Value Priced!
Download The World's Best XML IDE!
Accelerate XML development with our award-winning XML IDE - Download a free trial today!
Subscribe in XML format