Hello Tom, Martijn and List,
Thank you for these precisions.
I checked the spec file used to build the RPM and compile PostgreSQL and
I ran the configure with the icc compiler by setting :
%configure CC=/opt/intel/cc/9.1.045/bin/icc CFLAGS='-O2 -pipe
-override_limits'
I was afraid that this I
am Wed, dem 13.12.2006, um 7:55:48 + mailte Richard Huxton folgendes:
> Roderick A. Anderson wrote:
> >Works great ... up to a point. I now need to change the users and
> >owners associated with all the tables without changing their permissions.
> >
> >Is there a way or hack to do mass chan
Hello List,
Last question on the compilation of PostgreSQL 8.2 with the intel
compiler icc.
I work with an IA-64 server with Red Hat Enterprise Linux 4 AS update 2
on which the Intel compiler icc 9.1.045 is available and I build
PostgreSQL 8.2 with this compiler.
To do that I generated RPM fro
hello,
please read this post and the comment for it, maybe
the postgresql may need an update at separating the binaries.
this was pointed out in the FreeBSD ports collection.
- Forwarded message from Palle Girgensohn <[EMAIL PROTECTED]> -
Date: Tue, 12 Dec 2006 11:52:00 +0100
From: Pall
I just get my mind crossed here:
I do a simple select with order on a text-type. In my opinion entries with
leading spaces should be either first (or maybe last) in the list.
But in my select the whitespace just seems to be ignored:
Note that the second row has a leading space and should imho
peter pilsl wrote:
I just get my mind crossed here:
I do a simple select with order on a text-type. In my opinion entries
with leading spaces should be either first (or maybe last) in the list.
But in my select the whitespace just seems to be ignored:
That'll be down to your locale settings
Hi,
May be someone could help me with this:
For some time now, I exercise the use of VIEWs to expose just the
features of TABLES a particular user is supposed to see/have.
I can see that with a VIEW, I can do prity mutch everything I can do
with a TABLE, so a VIEW mimics a TABLE quite well b
Вот утром смотрю - сделался индекс (не знаю ск-ко часов он делался, но
тем не менее). То есть ошибок-то нету, в логе ругательств не
наблюдается.
Но вопрос остаётся.
Вопрос решился.
Дело было в системных лимитах на разделяемую память и семафоры, а
возможно и в размере кэша файловой системы. Опред
Sorry, wrong list.
> Вот утром смотрю - сделался индекс (не знаю ск-ко часов он делался, но
> тем не менее). То есть ошибок-то нету, в логе ругательств не
> наблюдается.
> Но вопрос остаётся.
Вопрос решился.
Дело было в системных лимитах на разделяемую память и семафоры, а
возможно и в размере
Richard Huxton wrote:
peter pilsl wrote:
I just get my mind crossed here:
I do a simple select with order on a text-type. In my opinion entries
with leading spaces should be either first (or maybe last) in the list.
But in my select the whitespace just seems to be ignored:
That'll be down
> Yeah, we could add defenses one by one for the cases we could think of,
> but I'd never feel very secure that we'd covered them all.
What you all forget in this discussion is that reindexing concurrently
would have to be a highly administrative task, controlled by the DB
admin... so whoever has
The problem here is that you've defined in business logic a relationship
between the records in your table, then *not defined that relationship
in the database*. Now you're complaining that the database doesn't do
what you want. How can it? You're not telling it everything.
Remember, in a relat
On mið, 2006-12-13 at 11:05 +0100, Csaba Nagy wrote:
> > Yeah, we could add defenses one by one for the cases we could think of,
> > but I'd never feel very secure that we'd covered them all.
>
> What you all forget in this discussion is that reindexing concurrently
> would have to be a highly adm
* Tom Lane:
> Florian Weimer <[EMAIL PROTECTED]> writes:
>> For regression tests, I'd like to automatically set up a fresh
>> PostgreSQL instance. Has anybody automated the task (initdb, setting
>> a password, choosing a port at random, starting the server, and after
>> running the tests, stoppin
On Tue, Dec 12, 2006 at 10:41:19PM -0500, Tom Lane wrote:
> "Brendan O'Shea" <[EMAIL PROTECTED]> writes:
> > Is there no way to specify a timeout for the write() to the socket or some
> > other way to abort?
>
> This is really a question to take up with your TCP stack implementors.
> I think it i
Tom Lane wrote:
Michael Fuhr <[EMAIL PROTECTED]> writes:
On Tue, Dec 12, 2006 at 12:19:56PM -0500, Tom Lane wrote:
Usually you do something like
select setval('seq_name', (select max(idcol) from table) + 1);
after loading data into the table.
Is "+ 1" necessary with the t
2006/9/28, Najib Abi Fadel <[EMAIL PROTECTED]>:
when u connect to the database type:
\h GRANT
and you will get all the Grant options:
GRANT { { CREATE | TEMPORARY | TEMP } [,...] | ALL [ PRIVILEGES ] }
ON DATABASE dbname [, ...]
TO { username | GROUP groupname | PUBLIC } [, ...] [ WITH G
On Wed, Dec 13, 2006 at 01:42:32PM +, Tomi N/A wrote:
> I don't get it. I grant all privileges on a database to a role, but
> the server won't let it access the schemas. I grant all privileges on
> the schema to the same role, but the server won't let it access the
> relations in the schema.
>
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 12/12/06 23:13, Anton wrote:
[snip]
>> Why do you have indexes on both LOGIN_ID *and* LOGIN_ID +
>> COLLECT_TIME?
> It is because I think that queries which use only LOGIN_ID field
> will use (faster) LOGIN_IDonly index... For me, speed of
> inserti
On 12/4/06, Bob Pawley <[EMAIL PROTECTED]> wrote:
Your missing the point.
I am creating a design system for industrial control.
The control devices need to be numbered. The numbers need to be sequential.
If the user deletes a device the numbers need to regenerate to again become
sequential and
Hi all,
I heard once that all messages which refer to table names should also
give their schema name.
I mean, if I see something like
rzeznia=# TRUNCATE nsp2.foo;
ERROR: cannot truncate a table referenced in a foreign key constraint
DETAIL: Table "bar" references "foo".
HINT: Truncate table
On 12/4/06, Bob Pawley <[EMAIL PROTECTED]> wrote:
Your missing the point.
I am creating a design system for industrial control.
The control devices need to be numbered. The numbers need to be
sequential.
If the user deletes a device the numbers need to regenerate to again
become
sequential
Hi -
I am struggling with a trigger function in plpgsql, and am hoping that
someone on this list can't show me a way to do what I need.
In the trigger, TG_ARGV[0] is the name of a column that I want to
evaluate. This code shows the concept, though is not functional:
CREATE OR REPLACE FUNCTI
Gergely CZUCZY <[EMAIL PROTECTED]> writes:
> This is wierd, I agree. The split between server and client is really made =
> in the postgresql sources,
> I just follow their guidelines.
[ checks RPM specfiles ... ] Both Devrim and I put pg_ctl in the server
subpackage, dunno where you think that t
Rafal Pietrak <[EMAIL PROTECTED]> writes:
> I can see that with a VIEW, I can do prity mutch everything I can do
> with a TABLE, so a VIEW mimics a TABLE quite well but one feature: a
> default value for a row on INSERT.
You can add a default to a view's column, either the same as the
underlyi
Marc Evans wrote:
Hi -
I am struggling with a trigger function in plpgsql, and am hoping that
someone on this list can't show me a way to do what I need.
In the trigger, TG_ARGV[0] is the name of a column that I want to
evaluate. This code shows the concept, though is not functional:
CREAT
Csaba Nagy <[EMAIL PROTECTED]> writes:
> So an implementation which optimistically builds the new index
> concurrently while holding no lock, and then hopes for the 3rd
> transaction to be able to get the exclusive lock and be able to swap the
> new index in the place of the old index, and error ou
On Wed, 13 Dec 2006, Erik Jones wrote:
Marc Evans wrote:
Hi -
I am struggling with a trigger function in plpgsql, and am hoping that
someone on this list can't show me a way to do what I need.
In the trigger, TG_ARGV[0] is the name of a column that I want to evaluate.
This code shows the
06-12-13, Shoaib Mir <[EMAIL PROTECTED]> napisał(a):
You can not truncate table 'foo' because there is referential integrity
between 'foo' and 'bar', so try using
TRUNCATE nsp2.foo CASCADE;
You can find the schema name for 'bar' using a query like:
select nspname from pg_namespace where oid =
Martijn van Oosterhout writes:
> I don't beleive you have to explicitly grant access to the database, or
> the schema, but you definitly have to grant access to the tables
> directly.
They're completely separate privileges. GRANT ON DATABASE grants or
revokes permissions associated with database
That would indeed be a handy feature... I was surprised when I discovered
this was not available like in mysql, it's a real pain to grant permissions
to a bunch of tables without querying table metadata.
GRANT ALL ON TABLE *.* TO joeblow
which would indeed be a useful thing to have, bu
Hi,
I have a user who is getting this error on a 8.1.3 server when importing
some data from a dbf file:
PostgreSQL Error Code: (1)
ERROR: date/time field value out of range: "08-01-2006"
He says the db is SQL-ASCII and the datestyle is at the default.
Of course on all my test servers the ins
Palle Girgensohn <[EMAIL PROTECTED]> writes:
> From INSTALL:
> ...
>Client-only installation: If you want to install only the client
>applications and interface libraries, then you can use these
>commands:
> gmake -C src/bin install
> gmake -C src/include install
> gmake -C
On Wed, 2006-12-13 at 17:12, Tom Lane wrote:
> Csaba Nagy <[EMAIL PROTECTED]> writes:
> > So an implementation which optimistically builds the new index
> > concurrently while holding no lock, and then hopes for the 3rd
> > transaction to be able to get the exclusive lock and be able to swap the
>
2006/12/13, Tom Lane <[EMAIL PROTECTED]>:
Martijn van Oosterhout writes:
> I don't beleive you have to explicitly grant access to the database, or
> the schema, but you definitly have to grant access to the tables
> directly.
They're completely separate privileges. GRANT ON DATABASE grants or
Marc Evans wrote:
On Wed, 13 Dec 2006, Erik Jones wrote:
Marc Evans wrote:
Hi -
I am struggling with a trigger function in plpgsql, and am hoping
that someone on this list can't show me a way to do what I need.
In the trigger, TG_ARGV[0] is the name of a column that I want to
evaluate. T
"=?UTF-8?Q?Filip_Rembia=C5=82kowski?=" <[EMAIL PROTECTED]> writes:
> I heard from my PostgreSQL teacher, that all such messages should be
> corrected in the source code, to include schema name too.
I doubt we'd ever do that, because in most scenarios it'd just be
unnecessary clutter, and the prima
--On onsdag, december 13, 2006 11.04.54 -0500 Tom Lane <[EMAIL PROTECTED]>
wrote:
Gergely CZUCZY <[EMAIL PROTECTED]> writes:
This is wierd, I agree. The split between server and client is really
made = in the postgresql sources,
I just follow their guidelines.
[ checks RPM specfiles ... ]
Greetings,
Could somebody shed any light on the error message below that came from
trying to call a simple pltcl test function? I am running Postgres 8.1
on WinXP and just recently added the pltcl language by copying Tcl84.dll
into my system directory (C:/Windows/System32) and successfully issui
http://developers.slashdot.org/article.pl?sid=06/12/13/1515217&from=rss
"MySQL quietly deprecated support for most Linux distributions on October 16,
when its 'MySQL Network' support plan was replaced by 'MySQL Enterprise.'
MySQL now supports only two Linux distributions — Red Hat Enterprise Lin
Marc Evans <[EMAIL PROTECTED]> writes:
> In the trigger, TG_ARGV[0] is the name of a column that I want to
> evaluate.
This is effectively impossible in plpgsql, because it's a statically
typed language --- it wants to know the type of every expression in
advance, and so such a thing couldn't wor
On Wed, 13 Dec 2006, Erik Jones wrote:
Marc Evans wrote:
On Wed, 13 Dec 2006, Erik Jones wrote:
Marc Evans wrote:
Hi -
I am struggling with a trigger function in plpgsql, and am hoping that
someone on this list can't show me a way to do what I need.
In the trigger, TG_ARGV[0] is the na
Tony Caduto <[EMAIL PROTECTED]> writes:
> I have a user who is getting this error on a 8.1.3 server when importing
> some data from a dbf file:
> ERROR: date/time field value out of range: "08-01-2006"
He needs to set DateStyle to match the expected field order.
regression=# show datestyle;
Da
David Goodenough wrote:
http://developers.slashdot.org/article.pl?sid=06/12/13/1515217&from=rss
"MySQL quietly deprecated support for most Linux distributions on October 16,
when its 'MySQL Network' support plan was replaced by 'MySQL Enterprise.'
MySQL now supports only two Linux distribution
On Wed, 2006-12-13 at 11:08 -0500, Tom Lane wrote:
>
> You can add a default to a view's column, either the same as the
> underlying table's default, or different if you want.
>
> ALTER TABLE view ALTER COLUMN col SET DEFAULT expr
G! The obvious solutions are most difficult to spot.
Thenx!
--
That means you might have the same table in both schemas, so try doing the
following to confirm:
select * from schema1.bar
select * from schema2.bar
You can also set search_path if you want to avoid adding schema name with
database objects:
set search_path = schema1
select * from bar
instead
Hi all,
I would like to know where postgres stores the created database and the
tables?Can anyone help?
thnx for nay help,
Antonios
-
Access over 1 million songs - Yahoo! Music Unlimited.
"Lenorovitz, Joel" <[EMAIL PROTECTED]> writes:
> Could somebody shed any light on the error message below that came from
> trying to call a simple pltcl test function? I am running Postgres 8.1
> on WinXP and just recently added the pltcl language by copying Tcl84.dll
> into my system directory (C
2006/12/13, Antonios Katsikadamos <[EMAIL PROTECTED]>:
I would like to know where postgres stores the created database and the
tables?
http://www.postgresql.org/docs/current/static/storage-file-layout.html
F.
---(end of broadcast)---
TIP 5: don't
On Wed, 2006-12-13 at 10:37 -0600, Tony Caduto wrote:
> Hi,
> I have a user who is getting this error on a 8.1.3 server when importing
> some data from a dbf file:
>
> PostgreSQL Error Code: (1)
> ERROR: date/time field value out of range: "08-01-2006"
>
> He says the db is SQL-ASCII and the da
Tom Lane wrote:
He needs to set DateStyle to match the expected field order.
regression=# show datestyle;
DateStyle
---
ISO, MDY
(1 row)
regression=# select '08-01-2006'::date;
date
2006-08-01
(1 row)
regression=# set datestyle TO ymd;
SET
regression=# select '08-01
Hi All,
This is something that bugs me for some time now. I have:
(as user postgres I do)
CREATE TABLE debi (id int, name text);
REVOKE ALL ON debi FROM public;
CREATE FUNCTION piti() RETURNS "trigger" AS $$ DECLARE me RECORD; BEGIN
select * into me FROM pg_authid; new.id
Merlin Moncure wrote:
On 12/13/06, Kevin Murphy <[EMAIL PROTECTED]> wrote:
I'd like to have a function that acts as a row filter (that can
optionally expand each row into multiple rows), but I don't know how to
wangle this such that the output is not enclosed in parentheses, i.e.
what I'm gettin
On Wed, 2006-12-13 at 10:50, David Goodenough wrote:
> http://developers.slashdot.org/article.pl?sid=06/12/13/1515217&from=rss
>
> "MySQL quietly deprecated support for most Linux distributions on October 16,
> when its 'MySQL Network' support plan was replaced by 'MySQL Enterprise.'
> MySQL now
Marc Evans wrote:
On Wed, 13 Dec 2006, Erik Jones wrote:
Marc Evans wrote:
On Wed, 13 Dec 2006, Erik Jones wrote:
Marc Evans wrote:
Hi -
I am struggling with a trigger function in plpgsql, and am hoping
that someone on this list can't show me a way to do what I need.
In the trigger, T
Scott Marlowe wrote:
> On Wed, 2006-12-13 at 10:50, David Goodenough wrote:
> > http://developers.slashdot.org/article.pl?sid=06/12/13/1515217&from=rss
> >
> > "MySQL quietly deprecated support for most Linux distributions on October
> > 16,
> > when its 'MySQL Network' support plan was replaced
On Wed, 2006-12-13 at 15:01 -0300, Alvaro Herrera wrote:
> Scott Marlowe wrote:
> > On Wed, 2006-12-13 at 10:50, David Goodenough wrote:
> > > http://developers.slashdot.org/article.pl?sid=06/12/13/1515217&from=rss
> > >
> > > "MySQL quietly deprecated support for most Linux distributions on Octob
On Wed, 2006-12-13 at 12:01, Alvaro Herrera wrote:
> Scott Marlowe wrote:
> > On Wed, 2006-12-13 at 10:50, David Goodenough wrote:
> > > http://developers.slashdot.org/article.pl?sid=06/12/13/1515217&from=rss
> > >
> > > "MySQL quietly deprecated support for most Linux distributions on October
>
The good thing is that there are several companies supporting
Postgres,
so whatever one of them does it does not affect the market as a whole.
Surely there are also third-party companies that provide "support"
for MySqueal in some similar sense?
- John Burger
MITRE
I'm wondering (given what I have read about PG and IN), if there is a
better way to execute this query. I have (essentially) two tables in
a read-only DB, ip_info and events, where many events can map to one
ip_info. I'd like to get all the ip_info.info columns where a
specific event occurred.
John D. Burger wrote:
The good thing is that there are several companies supporting Postgres,
so whatever one of them does it does not affect the market as a whole.
Surely there are also third-party companies that provide "support" for
MySqueal in some similar sense?
There probably are, b
* John D. Burger ([EMAIL PROTECTED]) wrote:
> >The good thing is that there are several companies supporting
> >Postgres,
> >so whatever one of them does it does not affect the market as a whole.
>
> Surely there are also third-party companies that provide "support"
> for MySqueal in some simi
On Wed, 2006-12-13 at 13:20 -0500, John D. Burger wrote:
> > The good thing is that there are several companies supporting
> > Postgres,
> > so whatever one of them does it does not affect the market as a whole.
>
> Surely there are also third-party companies that provide "support"
> for MySqu
"Lenorovitz, Joel" <[EMAIL PROTECTED]> writes:
> Could somebody shed any light on the error message below that came from
> trying to call a simple pltcl test function? I am running Postgres 8.1
> on WinXP and just recently added the pltcl language by copying Tcl84.dll
> into my system directory (C
I'm seeing some strange behavior with the following code. It compiles and
runs but returns an error on the Execute statement:
List index out of bounds(0)
DECLARE
srec record;
v_formula varchar;
v_result varchar;
BEGIN
v_formula = 'select 4 as val';
EXECUTE v_formula INTO srec;
END;
H
In response to "John D. Burger" <[EMAIL PROTECTED]>:
> > The good thing is that there are several companies supporting
> > Postgres,
> > so whatever one of them does it does not affect the market as a whole.
>
> Surely there are also third-party companies that provide "support"
> for MySqueal
Joshua D. Drake wrote:
Surely there are also third-party companies that provide "support"
for MySqueal in some similar sense?
Of course :) but... Fortune 2500+ for the most part will *not* use a
third party for support for something like MySQL.
Sure, but they won't use PG either, for essenti
Hi -
Is it accurate for me to believe that there is no way to catch exception
within plperl? For example, if an update fails, I'd like to do an insert.
Yes, this is trivial to do in plpgsql, but sadly plpgsql is not usable due
to not supporting variable variables. If plperl can't catch excepti
"Jon Asher" <[EMAIL PROTECTED]> writes:
> I'm seeing some strange behavior with the following code. It compiles and
> runs but returns an error on the Execute statement:
> List index out of bounds(0)
Worksforme. Want to provide a complete example instead of a fragmentary
one?
regression$# creat
John D. Burger wrote:
Sure, but they won't use PG either, for essentially the same reason,
since =all= PG support is "third party".
They would probably use EnterpriseDB though :-)
--
Tony Caduto
AM Software Design
http://www.amsoftwaredesign.com
Home of PG Lightning Admin for Postgresql
Yo
Rafal Pietrak <[EMAIL PROTECTED]> writes:
> I thought trigger functions execute at root/postgres security level?
No. You probably want to make that function SECURITY DEFINER so it
executes as the owner, but this isn't default for triggers.
regards, tom lane
-
Hi,
I'm using the function below to convert a large object to a bytea value.
Unfortunately, it performs very poorly for *large* objects, i.e. I'm
currently fiddling with a file of about 100 MB.
I've increased work_mem to 256 MB, but that didn't help much. I suspect
the aggregate function is
On Wed, 2006-12-13 at 13:00 -0600, Tony Caduto wrote:
> John D. Burger wrote:
> >
> > Sure, but they won't use PG either, for essentially the same reason,
> > since =all= PG support is "third party".
> >
> >
> They would probably use EnterpriseDB though :-)
Or Command Prompt like several extremel
Hi,
John D. Burger wrote:
Sure, but they won't use PG either, for essentially the same reason,
since =all= PG support is "third party".
Maybe. But at least these third parties can take the source and build
their own product on top of it, without significant limitations.
So one can debate if
On Wed, 2006-12-13 at 13:37 -0500, Bill Moran wrote:
> In response to "John D. Burger" <[EMAIL PROTECTED]>:
>
> > > The good thing is that there are several companies supporting
> > > Postgres,
> > > so whatever one of them does it does not affect the market as a whole.
> >
> > Surely there are
"Worky Workerson" <[EMAIL PROTECTED]> writes:
> I'm wondering (given what I have read about PG and IN), if there is a
> better way to execute this query.
> ... (as I have partitioned the events table by ip):
Are you on 8.2? 8.1 is pretty stupid about joins to inheritance trees.
Marc Evans <[EMAIL PROTECTED]> writes:
> Is it accurate for me to believe that there is no way to catch exception
> within plperl?
You do it the same way you trap any other error in perl ...
regards, tom lane
---(end of broadcast)-
Andreas, Richard,
Thanks. The link below is like of what I was looking for if there
wasn't an already written script. been through all three of my Pg books
but didn't find any mention of this type of scripting.
Rod
--
A. Kretschmer wrote:
am Wed, dem 13.12.2006, um 7:55:48 + mailte
Markus Schiltknecht <[EMAIL PROTECTED]> writes:
> John D. Burger wrote:
>> Sure, but they won't use PG either, for essentially the same reason,
>> since =all= PG support is "third party".
> So one can debate if i.e. EnterpriseDB is providing third party support
> for PostgreSQL or first-hand sup
i'm working with an FC5 test server which i'm trying to get set up
so we can benchmark xfs vs. jfs vs. ext3(writeback).
does anyone have any suggestions on setting up a tablespace on
a jfs partition with selinux in enforcing mode?
(i'm likely going to turn selinux off for the moment, but it'd
be
Tom Lane wrote:
The other point I'd make against John's argument is that there are a
whole lot of Fortune 500 companies buying Red Hat support, and RH is
effectively a third party for large chunks of Linux. (Of course,
there are also large chunks for which Red Hat employees write as much
code a
grant all privileges on database test to auser
As far as I can tell this does nothing. Intuitively this command suggests
that the auser would be able to access and modify the database test in
anyway. It would also suggest that as new tables for the database auser
would automatically have access
On Wed, Dec 13, 2006 at 12:20:03PM -0800, [EMAIL PROTECTED] wrote:
> grant all privileges on database test to auser
>
> As far as I can tell this does nothing. Intuitively this command suggests
> that the auser would be able to access and modify the database test in
> anyway. It would also sugge
Joshua D. Drake wrote:
> On Wed, 2006-12-13 at 13:20 -0500, John D. Burger wrote:
>> Surely there are also third-party companies that provide "support"
>> for MySqueal in some similar sense?
Yeah. HP for example [links below]. HP announced support
for Debian and MySQL (and the JBoss Stack as w
On Wed, 13 Dec 2006, Tom Lane wrote:
Marc Evans <[EMAIL PROTECTED]> writes:
Is it accurate for me to believe that there is no way to catch exception
within plperl?
You do it the same way you trap any other error in perl ...
OK, I must be missing something obvious:
c3i=> CREATE OR REPLACE
Marc Evans wrote:
OK, I must be missing something obvious:
c3i=> CREATE OR REPLACE FUNCTION foo_trigger() RETURNS TRIGGER AS $$
c3i$> eval { spi_exec_query('INSERT INTO FOO_BAR VALUES(1)'); } ||
c3i$> spi_exec_query('SELECT 1 as foo');
c3i$> $$ LANGUAGE plperl;
ERROR: creation of Perl funct
[EMAIL PROTECTED] wrote:
grant all privileges on database test to auser
As far as I can tell this does nothing. Intuitively this command suggests
that the auser would be able to access and modify the database test in
anyway. It would also suggest that as new tables for the database auser
would
I'm using an ancient version of postgresql (7.3 - don't ask) and I'm
experiencing a problem where many inserts into an empty table slow down
over time, even if I analyze in the middle of the insertions.
pg_stat_user_tables shows lots and lots of full scans, which explains the
slowdown, but lead
On Wed, Dec 13, 2006 at 02:08:44PM -0500, Tom Lane wrote:
> Are you on 8.2? 8.1 is pretty stupid about joins to inheritance trees.
To put up a number for "stupid": For GNUmed a particular
query involving one-level inheritance went down from over
ten minutes to under one second just by running ag
On Wed, Dec 13, 2006 at 01:10:44PM -0800, Ben wrote:
> I'm using an ancient version of postgresql (7.3 - don't ask) and I'm
> experiencing a problem where many inserts into an empty table slow down
> over time, even if I analyze in the middle of the insertions.
> pg_stat_user_tables shows lots a
On Wed, 2006-12-13 at 14:01 -0500, Tom Lane wrote:
> Rafal Pietrak <[EMAIL PROTECTED]> writes:
> > I thought trigger functions execute at root/postgres security level?
>
> No. You probably want to make that function SECURITY DEFINER so it
> executes as the owner, but this isn't default for trigge
On Wed, 2006-12-13 at 15:36, Rafal Pietrak wrote:
> On Wed, 2006-12-13 at 14:01 -0500, Tom Lane wrote:
> > Rafal Pietrak <[EMAIL PROTECTED]> writes:
> > > I thought trigger functions execute at root/postgres security level?
> >
> > No. You probably want to make that function SECURITY DEFINER so i
This is just an odd question. Previously RDBMSs (MySQL, SQLite, IIRC) I've
used have allowed row value constructors (optional feature F641 in SQL'03) on
INSERT statements. That is:
INSERT INTO mytable
VALUES (0,'hello'),(1,'world');
Is essentially shorthand for:
INSERT INTO mytable VALUES (0
On Wed, Dec 13, 2006 at 04:55:00PM -0500, Brandon Aiken wrote:
> This is just an odd question. Previously RDBMSs (MySQL, SQLite, IIRC) I've
> used have allowed row value constructors (optional feature F641 in SQL'03) on
> INSERT statements. That is:
>
> INSERT INTO mytable
> VALUES (0,'hello')
On Wed, 13 Dec 2006, Martijn van Oosterhout wrote:
- 7.3 isn't smart enough to use an index on an insert? Seems unlikely.
This question makes no sense, you don't need an index to insert.
Wouldn't it need to check the unique constraint (an index on the table)
before the insert can succeed?
Bricklen Anderson <[EMAIL PROTECTED]> writes:
> Marc Evans wrote:
>> OK, I must be missing something obvious:
>> ERROR: creation of Perl function failed: 'eval "string"' trapped by
>> operation mask at line 2.
> Try as plperlu
This brings up the question of whether it'd be safe to allow eval in
On Wed, Dec 13, 2006 at 02:01:46PM -0800, Ben wrote:
>
>
> On Wed, 13 Dec 2006, Martijn van Oosterhout wrote:
>
> >>- 7.3 isn't smart enough to use an index on an insert? Seems unlikely.
> >
> >This question makes no sense, you don't need an index to insert.
>
> Wouldn't it need to check the un
On Wed, Dec 13, 2006 at 05:04:42PM -0500, Tom Lane wrote:
> Bricklen Anderson <[EMAIL PROTECTED]> writes:
> > Marc Evans wrote:
> >> OK, I must be missing something obvious:
> >> ERROR: creation of Perl function failed: 'eval "string"' trapped by
> >> operation mask at line 2.
>
> > Try as plper
When you insert a tuple, it needs to be inserted into the index, yes. There
is no way an insert can cause a sequential scan, except by some trigger
defined on the table.
Actually, as it happens, there *is* a trigger defined on the table to fire
before insert, but it too uses an index scan, at l
I'd like to calculate the name of the WAL file associated with the
latest checkpoint using the output from pg_controldata:
Latest checkpoint location: 74/FD9827C0
Prior checkpoint location:74/FD6FCDC8
Latest checkpoint's REDO location:74/FD87DDF8
Latest checkpoint's UND
1 - 100 of 120 matches
Mail list logo