Subject: RE: Marked up data with sectbreaks (was Re: [xsl] XSLT FAQ ideas) From: "Passin, Tom" <tpassin@xxxxxxxxxxxx> Date: Tue, 18 Feb 2003 10:48:16 -0500 |
[ Elliotte Rusty Harold] > At 6:25 PM -0500 2/17/03, Wendell Piez wrote: > > >So far's I know that's a new solution or one not seen in > public anyhow. > > Cool. I invented something! In case anyone wants to see details look > at <http://www.cafeconleche.org/cafe.xsl> > > >The other known solutions are (1) indexing nodes to sectbreaks > >(generate-id(preceding::sectbreak[1]) works pretty well), then > >pulling them when matching sectbreaks ... documented somewhat in the > >FAQ under "flat" (flat structure to hierarchy). An alternative is a > >forward stepwise tree walk, in which a template matching a sectbreak > >has you proceed through the following sibling nodes one by one until > >you get a new sectbreak. > > > >The problem gets more complicated (much more) if your breaks are not > >all at the same level. > > > Yes, I can see that. Fortunately, mine are at the same level. I > wonder if multi-level breaking could be handled just by using the > following axis instead of following-sibling? > -- I had a nasty case where I had to extract some specific data items from an HTML page. There were markers before and after the data but they were different markers (different from each other, I mean) and in some cases at different levels. Once I could identify the right marker nodes and the candidate target data items, I took the intersection of all the candidate data items __after__ the first marker and __before__ the last marker. I forget now which of the axes I used for this particular case, following, following-sibling, etc. What made it worse was that the HTML (which I had no control over) was hand-written, invalid, and from time to time changed the way in which it was invalid. So I had to supply alternate expressions to make it work under the various observed permutations. (BTW, I used Tidy fo convert the HTML into usable xml). One example of this is that one of the end markers was an <a> element that had a <p> inside it (invalid!), and the <p> had an identifiable string I used to identify that it was indeed the marker. Sometimes the nesting of the a and p elements was reversed. Actually, the first set of items were used to index to the second set (they were url fragment identifiers pointing into the same page), and the second set actually pointed to the third set. The data I needed was a combination of the second and third sets. All three sets had to be extracted using thir own before and after markers. Whew! Cheers, Tom P XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
Re: [xsl] [XSD] Possible to ignore , Jeni Tennison | Thread | [xsl] incremental re-transformation, Felix Breuer |
Re: [xsl] [XSD] Possible to ignore , Jeni Tennison | Date | RE: [xsl] Weird problem. Xalan Stop, Passin, Tom |
Month |