James,

On 9 Jul 2003 at 10:32, James Cummings wrote:

> Ok.  Basically, I have one huge file that acts as a repository, which
> I then burst out into individual files so that they can each be
> accessed much quicker than pulling an individual element (by id()) out
> of the huge file.  I don't want to really get into putting the files
> into a database or anything along that line.  I'm happy to just do the
> bursting on the commandline using saxon, it will be a regular but
> probably infrequent event (a couple times a month at most), but it
> occurred to me that it might be good to have a map:match which would
> do this for me.  So I (or later someone) would have to just put the
> replacement repository in, and click said link to have all the files
> created. Just pure laziness, of course, and I could easily just do a
> quick .cgi of some sort to run the command.
> 
> -James
> 
> >
> >Regards, Upayavira
> >
> >On 9 Jul 2003 at 9:39, James Cummings wrote:
> >
> >> Hiya,
> >>
> >> Does the version of Xalan included with cocoon 2.1m3-dev allow XSLT
> >> 2.0?  Specifically does it allow xsl:result-document?
> >>
> >> Having one large file, I have an xslt stylesheet which bursts
> >> it into many individual files.  It works on the command line
> >> with saxon, but when I run the same stylesheet by doing:
> >>
> >>    <map:match pattern="burst-repository">
> >>    <map:generate src="repository/repository.xml"/>
> >>    <map:transform type="xalan" src="style/xsl/burst-repos.xsl">
> >>    <map:parameter name="use-request-parameters" value="true"/>
> >>    </map:transform> <map:serialize type="html"/> </map:match>
> >>
> >> The files are not created, and even worse, they appear to be
> >> appended to the output.
> >>
> >> I'm sure there is a way around this, and just something silly I've
> >> done.  Any suggestions?

Simplest way is to use the SourceWritingTransformer (SWT). You can have a 
pipeline that transforms your big XML file into a single XML that contains a number of 
nodes for the SWT - each one instructs it to create a file. Then, you transform the 
results of the SWT into some kind of HTML report explaining what happened.

You could wrap this pipeline in some kind of host selector/matcher, so that it can 
only 
be activated from your machine.

Then you can have pipelines that use your smaller files.

Make sense?

Regards, Upayavira

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to