Search solution for multiple databases

Forum Moderators: phranque

Message Too Old, No Replies

Search solution for multiple databases

what's a brotha to do :-)

walkman

2:52 pm on Jul 7, 2007 (gmt 0)

OK, I will have many databases: a yellow pages one, directory, news, forums, etc. etc... I would love to have a "Search" box that searches them all and spits out the results. Should I place a customized Google one or is there an internal one that does the job?

The problem with google is that I lose them to Google; I'd rather--obviously--keep them in my site template. Also, it woudl be great if it said:

Yellow Pages
-------
results....

Forums
-------
results

etc. etc.

any suggestions?

bill

1:02 am on Jul 8, 2007 (gmt 0)

I'd rather--obviously--keep them in my site template.

Google Co-op [google.com] lets you "Customize the look and feel to match your website."

peterdaly

1:16 am on Jul 8, 2007 (gmt 0)

Depending on your technical ability and resources, look into Nutch, SOLR, and Lucene. They are apache projects based in Java. Generally speaking you will need a dedicated server to run them.

Lucene is the core of a search engine, which you would need to write infrasture around, but what you write would fit your needs exactly.

Nutch is a full fledged web search engine (ala Google) using Lucene as the core.

SOLR is somewhat in-between. It's a web services interface to a fairly function search engine backend, again, based on Lucene.

All three options may be more technical than you are looking to get into, but are "professional grade" do it yourself solutions. Nutch would be the easiest to implement, but would be a cookie cutter type thing. SOLR or Lucene could be a fully custom search.

FWIW - I believe technorati is powered by Lucene. SOLR was originally written by cnet to power their product reviews engine. Yahoo has dabbled with supporting Nutch on an R&D level.

walkman

6:31 pm on Jul 8, 2007 (gmt 0)

thanks Peter and Bill.
The google thing looks good enough, and I guess it cuould work.

Is there a lot of server load involved with those other custom search engines? A setting them up I guess can be done with a few days of work or a freelancer.