Reopening a six months old discussion... Is there any plans to improve
ts_rank performance by making it able to use GIN/GIST indices?
>From re-reading Oleg answers, I understand FTS indices do not contain
enough information for ts_rank. Because of that, ts_rank has to read the
ts_vector, which is
On Wed, Jul 13, 2011 at 17:36, Oleg Bartunov wrote:
> I didn't notice, reading 40K tuples in random order takes a long time and
> this
> is a problem of any database. Can you measure time to read all documents
> found ?
As you asked, I measured the time required to read all documents.
For refere
I didn't notice, reading 40K tuples in random order takes a long time and this
is a problem of any database. Can you measure time to read all documents found ?
:( The only solution I see is to store enough
information for ranking in index.
Oleg
On Wed, 13 Jul 2011, Nicolas Grilly wrote:
Hi Ole
Hi Oleg and all,
On Wed, Jul 13, 2011 at 08:16, Oleg Bartunov wrote:
> there is problem with estimating of cost scanning gin index in < 9.1
> versions,
> so you can set enable_seqscan=off;
> or try 9.1 which beta3 now.
I re-ran my queries using enable seqscan=off.
Now the first query, without t
On Tue, 12 Jul 2011, Nicolas Grilly wrote:
On Tue, Jul 12, 2011 at 22:25, Oleg Bartunov wrote:
I don't see your query uses index :)
Yes, I know. :)
I ran VACUUM ANALYZE and re-ran the query but the output of EXPLAIN
ANALYZE stays exactly the same: no index used.
Any idea why?
there is pr
On Tue, Jul 12, 2011 at 22:25, Oleg Bartunov wrote:
> I don't see your query uses index :)
Yes, I know. :)
I ran VACUUM ANALYZE and re-ran the query but the output of EXPLAIN
ANALYZE stays exactly the same: no index used.
Any idea why?
By the way, does ts_rank is supposed to use a GIN index wh
I don't see your query uses index :)
On Tue, 12 Jul 2011, Nicolas Grilly wrote:
Hello,
I'm testing PostgreSQL full-text search on a table containing
1.000.000 documents. Document average length is 5.700 chars.
Performance is good and very similar to what I can get with Xapian if
I don't use ts
Hello,
I'm testing PostgreSQL full-text search on a table containing
1.000.000 documents. Document average length is 5.700 chars.
Performance is good and very similar to what I can get with Xapian if
I don't use ts_rank. But response time collapses if I use ts_rank to
select the 50 best matching d
