Re: Full text search?

[PA <petite.abeille@...>]

On Feb 28, 2005, at 21:20, Javier Guerra wrote:

> most people thought of something like this:

Yes, well... I contemplated using a straight table as well... but... 
decided against it as this doesn't help much for partial match... you 
still need to do a string.find() for that... or have a proper index 
structure... which I was too lazy to implement :P

Another consideration for using the document id as keys is that it much 
easier to update the index... just drop the entry for a document and 
reindex its text content... using an inverted index is too much house 
keeping for my humble needs.

Storing the text as a digest of sort slightly boost the string.find() 
performance (less stuff to scan) and provide an implicit ranking as 
more frequent tokens are found earlier in the digest.

All in all, kind of work ok for its simple purpose.

Cheers

--
PA, Onnay Equitursay
http://alt.textdrive.com/