On Wed, Jun 02, 2004 at 01:40:31PM -0700, Brad Miele wrote:
> Given a keyword, return a list of all keywords(non stemmed) in the index
> that are in documents which are matches for that word. I am tempted to
> just hit MySQL, but i think it might be faster to get from swish-e.
Index something big and then try it with both swish and with MySQL.
So, ignoring metanames and word positions?
There's nothing in swish that will do that automatically. You could
hack at the C code -- which would likely be the best way.
Let me get this clear:
1) search for "keyword" and get a list of files (actually file numbers)
that return results. What swish-e normally does:
2) For each file number:
search through the index and file every word that has that file
number and print it
So, use -T index_words and build yourself a filenumber -> list of words
Received on Wed Jun 2 14:19:07 2004