Re: 30-70 seconds query...

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "alexandre :: aldeia digital" <alepaes(at)aldeiadigital(dot)com(dot)br>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: 30-70 seconds query...
Date: 2003-03-31 22:17:20
Message-ID: 25610.1049149040@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

"alexandre :: aldeia digital" <alepaes(at)aldeiadigital(dot)com(dot)br> writes:
> I use a case tool and we generate the querys automatically.

> explain analyze SELECT T2.fi08ufemp, T4.es10almtra, T3.fi08MovEst,
> T1.es10qtdgra, T1.es10Tamanh, T1.es10item, T1.es10numdoc, T1.fi08codigo,
> T1.es10tipdoc, T1.es10codemp, T4.es10codalm, T4.es10empa, T1.es10datlan,
> T4.co13CodPro, T4.co13Emp06, T1.es10EmpTam FROM (((ES10T2 T1 LEFT JOIN
> ES10T T2 ON T2.es10codemp = T1.es10codemp AND T2.es10datlan =
> T1.es10datlan AND T2.es10tipdoc = T1.es10tipdoc AND T2.fi08codigo =
> T1.fi08codigo AND T2.es10numdoc = T1.es10numdoc) LEFT JOIN FI08T T3 ON
> T3.fi08ufemp = T2.fi08ufemp AND T3.fi08codigo =T1.fi08codigo) LEFT JOIN
> ES10T1 T4 ON T4.es10codemp = T1.es10codemp AND T4.es10datlan =
> T1.es10datlan AND T4.es10tipdoc = T1.es10tipdoc AND T4.fi08codigo =
> T1.fi08codigo AND T4.es10numdoc = T1.es10numdoc AND T4.es10item =
> T1.es10item) WHERE ( T4.co13Emp06 = '1' AND T4.co13CodPro = '16998' AND
> T1.es10datlan >= '2003-02-01'::date ) AND ( T1.es10datlan >=
> '2003-02-01'::date) AND ( T3.fi08MovEst = 'S' ) AND ( T4.es10empa = '1' OR
> ( '1' = 0 ) ) AND ( T4.es10codalm = '0' OR T4.es10almtra = '0' OR ( '0'
> = 0 ) ) AND ( T1.es10datlan <= '2003-02-28'::date ) ORDER BY
> T4.co13Emp06, T4.co13CodPro, T1.es10datlan, T4.es10empa, T4.es10codalm,
> T4.es10almtra, T1.es10codemp, T1.es10tipdoc, T1.fi08codigo,
> T1.es10numdoc, T1.es10item;

Your CASE tool isn't doing you any favors, is it :-(.

Mostly you need to rearrange the JOIN order into something more efficient.
I'd guess that joining T1 to T4, then to T3, then to T2 would be the
way to go here. Also, some study of the WHERE conditions proves that
all the LEFT JOINs could be reduced to plain joins, because any
null-extended row will get discarded by WHERE anyway. That would be a
good thing to do to give the planner more flexibility.

PG 7.4 will be better prepared to handle this sort of query, but I don't
think it will realize that the T1/T2 left join could be reduced to a
plain join given these conditions (that requires observing that null T2
will lead to null T3 because of the join condition... hmmm, I wonder how
practical that would be...). Without that deduction, the key step of
deciding to join T1/T4 first isn't reachable.

regards, tom lane

In response to

Responses

Browse pgsql-performance by date

  From Date Subject
Next Message Tom Lane 2003-03-31 22:58:32 Re: 30-70 seconds query...
Previous Message Tomasz Myrta 2003-03-31 22:15:30 Re: 30-70 seconds query...