Hi,
Tom Lane wrote:
> I've committed this change with (for now) 100 as the minimum histogram
> size to use. Stefan, are you interested in retrying your benchmark?
A first try with ltree gave big improvements on my smaller data set: the
estimated row count is correct or off by only 1 row. I'm now restoring a
bigger database to get more reliable results.
I hope Stefan can confirm the improvement on dbt3 too.
Thanks Tom :)
Best regards
--
Matteo Beccati
http://phpadsnew.com
http://phppgads.com