[opencms-dev] Best way to implement searching across article metadata

[Jonathan Woods]

Dear list...
 
I need to develop an OpenCms-based site which will allow users to search for
articles with matching metadata.  Some of this metadata is textual and
therefore ready for a Lucene-based search, but much of it is strongly typed
- e.g. dates, article importance, and references to other reference/standing
data which it would be a shame to model merely as text strings.  By the way,
the dates I mention are not simply dates already meaningful to OpenCms (e.g.
article creation date).  An example search might be 'show all articles with
--this-- subject concerning a date between --these-- dates'.
 
Could someone tell me roughly the best way to go about modelling articles to
make this kind of searching possible?  Is this something which the Lucene
module can cope with, provided that articles and their metadata are hooked
up to the Lucene index correctly?  Would I model this metadata using OpenCms
template properties (which I believe are merely strings), or as XML document
elements, or (God forbid!) in a separately maintained database structure?
I'm concerned not only about the most natural way to do this such that
content can be managed within OpenCms, but also about search performance - I
imagine searching through XML documents, for example, would be slow.
 
As if this weren't enough, I'm also wondering how best to mix metadata
searching like this with free-text searching (i.e. 'all articles on --this--
subject within --this date range containing these words').  If Lucene is the
answer above, then I can see how this can be done, but if not then the
picture isn't yet as clear.
 
Jon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://webmail.opencms.org/pipermail/opencms-dev/attachments/20060104/32fa801e/attachment.htm>