[XSL-LIST Mailing List Archive Home]
[By Thread]
[By Date]
Re: Optimisation (was Re[6]: Aggregate)
Subject: Re: Optimisation (was Re[6]: Aggregate) From: David Carlisle <davidc@xxxxxxxxx> Date: Fri, 17 Nov 2000 10:12:13 GMT |
> <xsl:number/> naively requires counting of the nodes that precede a given > node, which requires O(n) operations; therefore numbering a sequence of n > nodes requires O(n^2) operations. Saxon in some cases (but not all) > remembers the last node and its number, so that numbering a sequence of n > nodes becomes O(n). xt presumably does something similar as this one bit me (on a real job for NAG, not just when chatting on this list). We had something that was using <xsl:number level="any"/> which was working OK but when trying the stylesheet on the real document we realised that actually in general we needed to omit certain elements from the numbering depending on an attribute value so it looked like the small change to add something like count="xxx[not(ancestor-or-self::*/@os='nt')]" to not count (in the unix version of the spec) anything that is in a section flagged as nt only. (Norm, this is docbook, cf current thread on dsssl list...). On the real document (several hundred pages) the time to process went from a two or three minutes to (? I killed the job after half an hour....) David _____________________________________________________________________ This message has been checked for all known viruses by Star Internet delivered through the MessageLabs Virus Control Centre. For further information visit http://www.star.net.uk/stats.asp XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
RE: Optimisation (was Re[6]: Aggreg, Kay Michael | Thread | Re: Optimisation (was Re[6]: Aggreg, Jeni Tennison |
Re: O(n) notation (and character pa, Jeni Tennison | Date | Splitting result into 2 or more doc, Zeljko Rajic |
Month |
Keywords