Re: [BUGS] BUG #3975: tsearch2 index should not bomb out of 1Mb limit

Tom Lane wrote:
> Bruce Momjian <bruce(at)momjian(dot)us> writes:
> > Tom Lane wrote:
> >> I don't think that follows. A tsearch index is lossy anyway, so there's
>
> > Uh, the index is lossy but I thought it was lossy in a way that just
> > required additional heap accesses, not lossy in that it doesn't index
> > everything.
>
> Sure it's lossy. It doesn't index stopwords, and it doesn't index the
> difference between various forms of a word (when the dictionaries reduce
> them to a common root).

Yes, but you specify the stop words and stemming rules --- it isn't like
it drops words that are out of your control.

> > I am concerned a 1mb limit is too low though. Exactly why can't we have
> > a higher limit? Is positional information that significant?
>
> That's pretty much exactly the point: it's not very significant, and it
> doesn't justify a total inability to index large documents.

Agreed. I think losing positional information after 1mb is acceptable.

> One thing we could do is index words that are past the limit but not
> store a position, or perhaps have the convention that the maximum
> position value means "somewhere past here".

Sure.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://postgres.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +