| From: | Jon Nelson <jnelson+pgsql(at)jamponi(dot)net> |
|---|---|
| To: | pgsql-performance(at)postgresql(dot)org |
| Subject: | do temporary tables have hint bits? |
| Date: | 2010-11-13 13:53:27 |
| Message-ID: | AANLkTinDoe6sNb6X+WT0vG+pC5BEGpaCE649xT2CK0hq@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
I was doing some testing with temporary tables using this sql:
begin;
select pg_sleep(30);
create temporary TABLE foo (a int, b int, c int, d text);
insert into foo SELECT (x%1000) AS a,(x%1001) AS b, (x % 650) as c, ''
as d FROM generate_series( 1, 1000000 ) AS x;
-- create temporary TABLE foo AS SELECT (x%1000) AS a,(x%1001) AS b,
(x % 650) as c, '' as d FROM generate_series( 1, 1000000 ) AS x;
select count(1) from foo;
While it was in pg_sleep, I would attach to the backend process with strace.
I observed a few things that I don't yet understand, but one thing I
did notice was an I/O pattern (following the count(1)) that seemed to
suggest that the table was getting its hint bits set. I thought hint
bits were just for the mvcc side of things? If this is a temporary
table, is there any need or benefit to setting hint bits?
--
Jon
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2010-11-13 14:57:29 | Re: do temporary tables have hint bits? |
| Previous Message | Marc Mamin | 2010-11-13 09:32:12 | Re: anti-join chosen even when slower than old plan |