pg.ddx.io pgsql-interfaces@postgresql.org mailing list archive
help / color / mirror / Atom feedPossible Bug regarding temp tables (sql or psql?)
11+ messages / 4 participants
[nested] [flat]
* Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-03 20:59 Mark Frazer <mark@somanetworks.com>
0 siblings, 1 reply; 11+ messages in thread
From: Mark Frazer @ 2002-07-03 20:59 UTC (permalink / raw)
To: pgsql-sql@postgresql.org; pgsql-interfaces@somanetworks.com
When using the attached script in psql, the temp variables disappear as
far as \distv shows, but running the script a second time fails.
To reproduce, save the following script as bug.sql, then start psql on an
test database.
\i bug.sql
\distv
-- no relations should be shown
\i bug.sql
-- this will fail.
However, if you run psql again, you can
\i bug.sql
successfully. It will only fail if run twice in the same script.
cheers
-mark
--
Hardy Boys: too easy. Nancy Drew: too hard! - Fry
Attachments:
bug.sql (text/plain, 536B) download
CREATE TABLE foo (
foo_idx SERIAL PRIMARY KEY,
foo INTEGER ) ;
CREATE TABLE bar (
bar_idx SERIAL PRIMARY KEY,
foo_idx INTEGER REFERENCES foo,
bar INTEGER ) ;
INSERT INTO foo ( foo ) VALUES ( 111 ) ;
INSERT INTO foo ( foo ) VALUES ( 222 ) ;
INSERT INTO foo ( foo ) VALUES ( 333 ) ;
BEGIN WORK ;
SELECT foo_idx INTO TEMP v_idx FROM foo WHERE foo.foo = 222 ;
INSERT INTO bar ( foo_idx, bar ) VALUES ( v_idx.foo_idx, 888 ) ;
END WORK ;
DROP TABLE foo ;
DROP SEQUENCE foo_foo_idx_seq ;
DROP TABLE bar ;
DROP SEQUENCE bar_bar_idx_seq ;
* Re: Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-03 21:12 Mark Frazer <mark@somanetworks.com>
parent: Mark Frazer <mark@somanetworks.com>
0 siblings, 1 reply; 11+ messages in thread
From: Mark Frazer @ 2002-07-03 21:12 UTC (permalink / raw)
To: pgsql-sql@postgresql.org; pgsql-interfaces
Forgot to mention that adding
DROP TABLE v_idx ;
before the END WORK will fix things. However, I was under the impression that
temporary tables would go away after a transaction in which they were created
was committed.
^ permalink raw reply [nested|flat] 11+ messages in thread
* Re: Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-04 01:38 Christopher Kings-Lynne <chriskl@familyhealth.com.au>
parent: Mark Frazer <mark@somanetworks.com>
0 siblings, 1 reply; 11+ messages in thread
From: Christopher Kings-Lynne @ 2002-07-04 01:38 UTC (permalink / raw)
To: Mark Frazer <mark@somanetworks.com>; pgsql-sql@postgresql.org
> Forgot to mention that adding
> DROP TABLE v_idx ;
> before the END WORK will fix things. However, I was under the
> impression that
> temporary tables would go away after a transaction in which they
> were created
> was committed.
No - they go away at the end of a _connection_. However, there is now a
patch floating around on -hackers that would add an ' ON COMMIT DROP;'
option to CREATE TEMP TABLE.
Chris
^ permalink raw reply [nested|flat] 11+ messages in thread
* Re: [SQL] Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-04 11:23 Mark Frazer <mark@somanetworks.com>
parent: Christopher Kings-Lynne <chriskl@familyhealth.com.au>
0 siblings, 2 replies; 11+ messages in thread
From: Mark Frazer @ 2002-07-04 11:23 UTC (permalink / raw)
To: pgsql-sql@postgresql.org; +Cc: pgsql-interfaces
Any idea why the table can't be seen with \d in psql then?
Christopher Kings-Lynne <chriskl@familyhealth.com.au> [02/07/04 00:21]:
>
> No - they go away at the end of a _connection_. However, there is now a
> patch floating around on -hackers that would add an ' ON COMMIT DROP;'
> option to CREATE TEMP TABLE.
--
In the event of an emergency, my ass can be used as a flotation
device. - Bender
^ permalink raw reply [nested|flat] 11+ messages in thread
* Re: [SQL] Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-04 12:26 Bruce Momjian <pgman@candle.pha.pa.us>
parent: Mark Frazer <mark@somanetworks.com>
1 sibling, 0 replies; 11+ messages in thread
From: Bruce Momjian @ 2002-07-04 12:26 UTC (permalink / raw)
To: Mark Frazer <mark@somanetworks.com>; +Cc: pgsql-sql@postgresql.org; pgsql-interfaces
Mark Frazer wrote:
> Any idea why the table can't be seen with \d in psql then?
It is a known problem with temp tables. They aren't visible with \d
because they are invisible system tables that are removed on exit. I
think 7.3 will fix this.
>
> Christopher Kings-Lynne <chriskl@familyhealth.com.au> [02/07/04 00:21]:
> >
> > No - they go away at the end of a _connection_. However, there is now a
> > patch floating around on -hackers that would add an ' ON COMMIT DROP;'
> > option to CREATE TEMP TABLE.
>
> --
> In the event of an emergency, my ass can be used as a flotation
> device. - Bender
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 3: if posting/reading through Usenet, please send an appropriate
> subscribe-nomail command to majordomo@postgresql.org so that your
> message can get through to the mailing list cleanly
>
>
>
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026
^ permalink raw reply [nested|flat] 11+ messages in thread
* Re: Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-05 02:02 Christopher Kings-Lynne <chriskl@familyhealth.com.au>
parent: Mark Frazer <mark@somanetworks.com>
1 sibling, 1 reply; 11+ messages in thread
From: Christopher Kings-Lynne @ 2002-07-05 02:02 UTC (permalink / raw)
To: Mark Frazer <mark@somanetworks.com>; pgsql-sql@postgresql.org
IIRC they are only visible in the connection that created them? Am I right?
Either way, I don't think \d will show them - there might be another
command...
Chris
> Any idea why the table can't be seen with \d in psql then?
>
> Christopher Kings-Lynne <chriskl@familyhealth.com.au> [02/07/04 00:21]:
> >
> > No - they go away at the end of a _connection_. However, there is now a
> > patch floating around on -hackers that would add an ' ON COMMIT DROP;'
> > option to CREATE TEMP TABLE.
^ permalink raw reply [nested|flat] 11+ messages in thread
* Re: Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-05 02:09 Bruce Momjian <pgman@candle.pha.pa.us>
parent: Christopher Kings-Lynne <chriskl@familyhealth.com.au>
0 siblings, 2 replies; 11+ messages in thread
From: Bruce Momjian @ 2002-07-05 02:09 UTC (permalink / raw)
To: Christopher Kings-Lynne <chriskl@familyhealth.com.au>; +Cc: Mark Frazer <mark@somanetworks.com>; pgsql-sql@postgresql.org
Christopher Kings-Lynne wrote:
> IIRC they are only visible in the connection that created them? Am I right?
> Either way, I don't think \d will show them - there might be another
> command...
>
TODO has:
* Allow psql \d to show temporary table structure
Looks like it works fine now with schemas:
test=> create temp table xx(x int);
CREATE TABLE
test=> \d
List of relations
Name | Type | Owner
------+-------+----------
x | table | postgres
xx | table | postgres
(2 rows)
I will mark the TODO as done.
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026
^ permalink raw reply [nested|flat] 11+ messages in thread
* Re: Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-05 02:32 Tom Lane <tgl@sss.pgh.pa.us>
parent: Bruce Momjian <pgman@candle.pha.pa.us>
1 sibling, 1 reply; 11+ messages in thread
From: Tom Lane @ 2002-07-05 02:32 UTC (permalink / raw)
To: Bruce Momjian <pgman@candle.pha.pa.us>; +Cc: Christopher Kings-Lynne <chriskl@familyhealth.com.au>; Mark Frazer <mark@somanetworks.com>; pgsql-sql@postgresql.org
Bruce Momjian <pgman@candle.pha.pa.us> writes:
> TODO has:
> * Allow psql \d to show temporary table structure
> Looks like it works fine now with schemas:
More accurately, it's no more broken for temp tables than any other
tables at the moment.
regression=# create table foo (f1 int);
CREATE TABLE
regression=# create temp table foo (f1 text, f2 float);
CREATE TABLE
regression=# \d foo
Table "foo"
Column | Type | Modifiers
--------+------------------+-----------
f1 | integer |
f1 | text |
f2 | double precision |
regression=#
Even uglier cases occur if some other connection has also created
a temp table named foo. You should *not* see that in \d ... but
you will.
> I will mark the TODO as done.
It's not really done, but the remaining fixes will fall out of making
psql schema-aware.
regards, tom lane
^ permalink raw reply [nested|flat] 11+ messages in thread
* Re: Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-05 02:39 Bruce Momjian <pgman@candle.pha.pa.us>
parent: Tom Lane <tgl@sss.pgh.pa.us>
0 siblings, 0 replies; 11+ messages in thread
From: Bruce Momjian @ 2002-07-05 02:39 UTC (permalink / raw)
To: Tom Lane <tgl@sss.pgh.pa.us>; +Cc: Christopher Kings-Lynne <chriskl@familyhealth.com.au>; Mark Frazer <mark@somanetworks.com>; pgsql-sql@postgresql.org
Tom Lane wrote:
> Even uglier cases occur if some other connection has also created
> a temp table named foo. You should *not* see that in \d ... but
> you will.
>
> > I will mark the TODO as done.
>
> It's not really done, but the remaining fixes will fall out of making
> psql schema-aware.
The original bug is gone. We just have new ones.
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026
^ permalink raw reply [nested|flat] 11+ messages in thread
* Re: Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-05 13:38 Mark Frazer <mark@somanetworks.com>
parent: Bruce Momjian <pgman@candle.pha.pa.us>
1 sibling, 1 reply; 11+ messages in thread
From: Mark Frazer @ 2002-07-05 13:38 UTC (permalink / raw)
To: Bruce Momjian <pgman@candle.pha.pa.us>; +Cc: Christopher Kings-Lynne <chriskl@familyhealth.com.au>; pgsql-sql@postgresql.org
Bruce Momjian <pgman@candle.pha.pa.us> [02/07/04 22:10]:
>
> TODO has:
>
> * Allow psql \d to show temporary table structure
>
> Looks like it works fine now with schemas:
>
> I will mark the TODO as done.
It doesn't work with select into though:
config=> select 5 into temp v_tmp ;
SELECT
config=> \d v_tmp
Did not find any relation named "v_tmp".
config=> select 4 into temp v_tmp ;
ERROR: Relation 'v_tmp' already exists
config=> select version() ;
version
-------------------------------------------------------------
PostgreSQL 7.2.1 on i686-pc-linux-gnu, compiled by GCC 2.96
--
I heard one time you single-handedly defeated a hoard of rampaging somethings
in the something something system. - Fry
^ permalink raw reply [nested|flat] 11+ messages in thread
* Re: Possible Bug regarding temp tables (sql or psql?)
@ 2002-07-05 18:52 Bruce Momjian <pgman@candle.pha.pa.us>
parent: Mark Frazer <mark@somanetworks.com>
0 siblings, 0 replies; 11+ messages in thread
From: Bruce Momjian @ 2002-07-05 18:52 UTC (permalink / raw)
To: Mark Frazer <mark@somanetworks.com>; +Cc: Christopher Kings-Lynne <chriskl@familyhealth.com.au>; pgsql-sql@postgresql.org
It works, but only in CVS, not in 7.2.X. Marking something as done in
TODO only means it is done and will be in the _next_ release. Sorry.
---------------------------------------------------------------------------
Mark Frazer wrote:
>
>
> Bruce Momjian <pgman@candle.pha.pa.us> [02/07/04 22:10]:
> >
> > TODO has:
> >
> > * Allow psql \d to show temporary table structure
> >
> > Looks like it works fine now with schemas:
> >
> > I will mark the TODO as done.
>
> It doesn't work with select into though:
>
> config=> select 5 into temp v_tmp ;
> SELECT
> config=> \d v_tmp
> Did not find any relation named "v_tmp".
> config=> select 4 into temp v_tmp ;
> ERROR: Relation 'v_tmp' already exists
> config=> select version() ;
> version
> -------------------------------------------------------------
> PostgreSQL 7.2.1 on i686-pc-linux-gnu, compiled by GCC 2.96
>
>
> --
> I heard one time you single-handedly defeated a hoard of rampaging somethings
> in the something something system. - Fry
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Don't 'kill -9' the postmaster
>
>
>
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026
^ permalink raw reply [nested|flat] 11+ messages in thread
end of thread, other threads:[~2002-07-05 18:52 UTC | newest]
Thread overview: 11+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2002-07-03 20:59 Possible Bug regarding temp tables (sql or psql?) Mark Frazer <mark@somanetworks.com>
2002-07-03 21:12 ` Mark Frazer <mark@somanetworks.com>
2002-07-04 01:38 ` Christopher Kings-Lynne <chriskl@familyhealth.com.au>
2002-07-04 11:23 ` Mark Frazer <mark@somanetworks.com>
2002-07-04 12:26 ` Bruce Momjian <pgman@candle.pha.pa.us>
2002-07-05 02:02 ` Christopher Kings-Lynne <chriskl@familyhealth.com.au>
2002-07-05 02:09 ` Bruce Momjian <pgman@candle.pha.pa.us>
2002-07-05 02:32 ` Tom Lane <tgl@sss.pgh.pa.us>
2002-07-05 02:39 ` Bruce Momjian <pgman@candle.pha.pa.us>
2002-07-05 13:38 ` Mark Frazer <mark@somanetworks.com>
2002-07-05 18:52 ` Bruce Momjian <pgman@candle.pha.pa.us>
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox