Re: general purpose full text indexing

----- Original Message -----
From: "Avishai Weissberg" <ovesh1(at)yahoo(dot)com>
To: <pgsql-general(at)postgresql(dot)org>
Sent: Monday, March 28, 2005 2:00 PM
Subject: [GENERAL] general purpose full text indexing

> Hello,
>
> I am trying to find a suitable FTI component.
>
> I am aware of tsearch2, but as far as I understand it doesn't really suit
> my purposes. I want to be able to run a search on a huge TEXT column,
> where the the column's content is made of words (each 'word' is an email
> address) seperated by white-spaces and commas. The search should also
> support wild-chars. Lexemes are not needed, as the 'words' are not in any
> language.
>
> An example entry would be:
> ===========
> 'anbd(at)ijadf(dot)com, okok(at)uhuh(dot)org,plpl(at)erer(dot)com, spsp(at)qaqa(dot)ne(dot)jp'
> ===========
> Entries run up to several thousand charaters.
>
> An example query would be:
> ===========
> WHERE column_name like '%(at)uhuh(dot)org' or column_name like '%plpl(at)erer(dot)com%'
> ===========
>
> Is there a way to do this using tsearch2 without implementing a new
> parser?
> Is there another component of which I am not aware for full text indexing?

This exact same question (searching email columns) has come up in the past
couple of weeks. I tried searching the archives for the reply, but it
appears that, at least for me, search.postgresql.org is down right now. In
any case, if I remember correctly, the answer was to look at:

http://sun.calstatela.edu/~cysun/documentation/postgres/8/contrib/README.pg_trgm

Hope this helps. You should also probably try searching the archives, as I
think there are answers there.

Sean