Subject: Re: [xsl] Pattern question: first child of first descendant From: Andrew Welch <andrew.j.welch@xxxxxxxxx> Date: Tue, 24 Mar 2009 11:08:21 +0000 |
>>There may well be something better, but I would probably have used: >> table//entry[1]/*[self::title or self::para][1] > > Me too, until I found that it is wrong :-) It does "match the first > title or para child of every entry descendant in the table that is the > first child of some other element". > > E.g. in the common case where entry elements are wrapped by row > elements, the above pattern matches every first para or title element in > the first entry child of each row in the table, not just the first entry > element (in document order) in the whole table. That's the old // gotcha, where you want to use the descendant axis instead... As in this case, it often in the form "//somenode[1]" - some good info on it can be found here: http://markmail.org/message/ocal22heg7wow3yi#query:related%3Aocal22heg7wow3yi +page:1+mid:q3detja3uzvunrn6+state:results -- Andrew Welch http://andrewjwelch.com Kernow: http://kernowforsaxon.sf.net/
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
Re: [xsl] Pattern question: first c, Christian Roth | Thread | Re: [xsl] Pattern question: first c, David Carlisle |
Re: [xsl] Pattern question: first c, David Carlisle | Date | Re: [xsl] Pattern question: first c, David Carlisle |
Month |