Re: [htdig] Multiple datasets combined.


Shantonu Sen (ssen@MIT.EDU)
Mon, 22 Mar 1999 12:03:45 -0500


ok, this may be a little far-fetched, but here's my initial thought.

now, this idea depends on how htsearch accesses header.html and
footer.html. if it essentially preprnds/appends the files, and then
substitutes in variables, this might work. i will be using the numers 1, 2,
and 3, but everything should be named according to standard conventions,
just in differnt directories.

my idea is to have a search form that passes the arguments to a shell
script. the shell script would in turn pass the arguments on to three
htsearch process, each of which would point to a different htdig.conf file,
and the results of all three would be concatenated

now, htdig1.conf would specify a normal header.html file, but it would
specify a _blank_ footer.html, because you dont want the HTML code to
terminate, because there are more hits on the way, presumably!

for a similar reason, header2.html and footer2.html would be blank, because
you odn't want dilineation between results.

finally, htdig3.conf would specify a blank header.html file, and the _real_
footer.html file.

to make this work, you'd also want to make the "no results found" page
blank, because you dont want that in the middle of a search result page.

you'd have to start the sheel script by printing out MIME-types, but you
can probably figure this out.

now, the fun part is making this script smart. you could save each of the
search runs to a different temporary file. if all returned no hits (you'll
be able to tell because the "no results" page is blank, so each of the
temporary files should be blank), then print out the _actual_ no results
found page.

also, this method will only let you print _all_ the hits from all the
databases. you could have the script intercept what "page of hits" and
"number of hits per page" the user is looking for, and depending on whether
any of the temporary files has size 0, rerun the the search to get more
hits from database that actually have files you want.

i realize this is a porr solution, because you will not be able to rank
across databases, or even intersperse their results, but if you want all
hits to show up, this may or may not work.

good luck!
shantonu sen

At 02:09 PM 3/22/99 -0000, you wrote:
>
>Is it possible for multiple databases to be combined for the purposes of a
>search?
>
>I don't want to *merge* the databases, I simply want to search across
>multiple databases.
>
>If it is not possible, and hacks, ideas?
>
>
>
>
>Barry Zubel
>Technical Manager
>City Mutual Ltd
>www.citymutual.com
>
>------------------------------------
>To unsubscribe from the htdig mailing list, send a message to
>htdig@htdig.org containing the single word "unsubscribe" in
>the SUBJECT of the message.
>
>
>
-----------------------------------------------------------
Shantonu Sen
Massachusetts Institute of Technology
ssen@mit.edu
------------------------------------
To unsubscribe from the htdig mailing list, send a message to
htdig@htdig.org containing the single word "unsubscribe" in
the SUBJECT of the message.



This archive was generated by hypermail 2.0b3 on Mon Mar 22 1999 - 14:06:37 PST