>>Basically you can setup a table structure in whatever language your
>>using to display your results, and do a lookup on that table based on
>>the filename..... E.g., using PHP....
> Although, I think in the general case one would want to store the
> title in the swish index instead of looking when returning results.
Absolutly, depending on the context of the data of course. I currently
do a very simmilar thing with data straight from a MYSql database (e.g.
products). I first convert it to a html file so it indexs fine, but when
I pull out the results I only extract the product id, which I then use
to call the database for the content of the results.
In this case it always displays the current price (which may have
changed since last index) and image and information etc. The only thing
thats sacraficed is the index file is possibly out of sync till the next
index. But more importantly, the results show the current price, not the
However, in the case of PDF files, I agree that the general cause should
be to index the title along with the pdf content.
Having said that - I just love the flexability of Swish-E to be able to
use whichever method you that works for a particular situation. My
previous experience was with an engine called IPhrase. Quite powerful,
but extremely difficult to display the results (the version that I was
using only had the xml/xslt option for displaying results). I had
swish-e installed in less than a week, now our results not only look
like they are on our site (IPhrase uses its own header etc, which on our
site is dynamic, so was always out of date) but also shows results that
are accurate (IPhrase indexed the price which was useful, but not if we
have a price increase - people would click on the search results and
ended up adding to our cart at a higher price).
Anyway, short story: I'm loving swish-e :)
Received on Wed Nov 23 19:50:59 2005