[rt-users] EXTERNAL: Re: RT Include substrings with Full-Text Indexing (PostgreSQL)

[ktm at rice.edu]

On Thu, Jul 26, 2012 at 07:26:32PM +0000, Betz, Gregory wrote:
> Hey Ken,
> 
> So right know everything is indexed and searching works correctly except that I cannot search on a substring.
> 
> For example, 
> 
> I have a ticket with the following contained within.
> https://encrypted.google.com/search?q=rt
> 
> 
> I search for "google.com"...
> No results are returned.
> 
> I search  for "https://encrypted.google.com/search?q=rt" ...
> The ticket with the content is returned.
>  
> It looks as though the content within ticket is indexed based off of whitespaces into the attachments table.
> 
> 
> Thanks,
> Greg
> 

Hi Greg,

Row I get it. I have had the same problem and it has been discussed on
the PostgreSQL mailing lists. This is caused by the parser/tokenizer that is
used by default in PostgreSQL. It only indexes the "full" hostname. There was
someone working to change some of these behaviors in the default parser. In
particular a domain/sub-domain part of a token would also be indexed. I do
not think that it has been finished and I have not seen any recent discussions.

Cheers,
Ken