Re: [HACKERS] postgres performance

Shiby Thomas wrote:
>
> =>      There may be optimizations in the 2.6 libraries that would improve
> => performance, but I wouldn't suspect that it would make *that* big of a
> => difference.  What is your SQL/join statemnt?  How are you running
> => postmaster?  What does 'explain' show?
> =>
> The complete query is this:
>
> select item1, item2, count(t1.tid) into table f2_temp from data t1, data t2,
> c2
> where t1.item = c2.item1 and t2.item = c2.item2 and t1.tid = t2.tid group by
> ite
> m1, item2
>
> data is a table with 2 integer columns (tid, item) and it has ~300K records
> c2 is a table (item1, item2), both integers and has ~1.5K records.
>
> I was directly running postgres with the -B and -S flags to give more buffers
> and sortMem. I also tried several join plans by the -f flags. Hash join works
> the best and that itself is too slow (perhaps due to the self join)

Indices ?
EXPLAIN ?

Vadim