Google Style Optimized Search of a Database

[MarillionFan]

If I want to know about Access or Excel, I'll definitely ask you Dim otherwise I'll wait for an answer from a proper contractor.

[scooterscot]

PS: Information systems is what you're wanting to look into, optimisation concerns minimising resource to generate a desired outcome.

[DimPrawn]

Originally Posted by MarillionFan

If I want to know about Access or Excel, I'll definitely ask you Dim otherwise I'll wait for an answer from a proper contractor.

Atw is on his way over to kick your butt.

[Bob Dalek]

Try Lucene. It's now Oracle-savvy, so give it a go.

[plinthadenoid]

I've been working with someone who took over a project that has so far taken two years of a data analyst time working up a taxonomy of the data to be searched to allow quick and easy searching on-line.

Chap I work with has just recommended the client stops trying to organise the data which they have little control over and instead plump for one of these:[url="http://www.google.com/enterprise/gsa/"]

Costs don't seem to be too bad, small box is £2k, enterprise approx £18k.

Would it be relevant in your case?

Plinth

[DimPrawn]

http://www.microsoft.com/enterprises...s/default.aspx

Free as in free beer.

[plinthadenoid]

Will SSE 2008 crawl an Oracle database?

[NickFitz]

Originally Posted by DimPrawn

http://www.microsoft.com/enterprises...s/default.aspx

Free as in free beer.

Is that what they use for the search on microsoft.com? Because that sucks...

[fatcat51]

+1 for Lucene, just integrated it into a .net project on sql server 2005. It's blazingly fast and really flexible.

I was shocked how quick and easy it was to get search up and running, plus gives you loads of extra features like highlighting etc.

[xchaotic]

In this particular example, I dare say that the db is not flat.
Which means you could arrange fields to search for a location first and have an index of those.

Also within those 55m there will be repetitions (companies with the same or a similar name), so you only need to hint at viable possibilities.
Notice how google only does that after you provide the first word, as that limits the number of possibilities.