Hi all,
I'm using psql 8.0.0 on a client's site who's running win server 2003.
We've had him on beta 3 for some time, and no problems at all
(yes, in a sense, he is a beta tester as well, but doesn't
know it!). Today I tried to upgrade the db to RC1 and had
some problems.
Remote
Hi
I am facing some issues with select query. The
values of the columns in one table contains \n as a
part of the value. So when I execute the select query
on this table, I am getting the following output.
# select * from stdhlr_subscriber_profile ;
subscriber_id | ssd_a
- Original Message -
From: Marc G. Fournier [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Sent: Wednesday, December 22, 2004 1:37 AM
Subject: [ANNOUNCE] PostgreSQL 8.0.0 Release Candidate 2
Its been 3 weeks since our first Release Candidate, and we're down to the
Hi all,
I have 2 tables, with a 1-n relation:
parent( oid, parent_name)
child(oid, child_name, iod_parent)
How do I get the parent_names of all parents without a child?
TIA
Groeten,
Joost Kraaijeveld
Askesis B.V.
Molukkenstraat 14
6524NB Nijmegen
tel: 024-3888063 / 06-51855277
fax:
Hi All!
I tried to find out more details about this weird problem, and I must
say, I am stumped.
here are the details:
We run a software module on a 64-bit opteron with a 32 Gig RAM, RedHat
AS3 update 3, Linux 2.4.21. postgres 7.4.6
I dumped the postgres query logs and here are the numbers:
Hi,
I want to turn off the default setting in postgres for index creation on
primary key of a table. Is it possible and how?
Regards
Vinita
_
Citibank Suvidha account at No Minimum Balance!
Hi all,
I have 2 tables, with a 1-n relation:
parent( oid, parent_name)
child(oid, child_name, iod_parent)
How do I get the parent_names of all parents without a child?
select parent_name from parent
left join child on (parent.oid=child.iod_parent)
where child.oid is null;
or
select parent_name
Hi,
Is it a normal behavior that if I give a where clause with an existent index
key, then postgres uses the index, but if I give it a non existent value
than it refuses to use the index.
An example to make it more clear:
CREATE TABLE measured_1
(
tstamp timestamp(0) NOT NULL,
meterid int4
The FSF site explicitly says that free software is not free, it's about
freedom. I don't think the idea is to let your kids starve. --Rick
Hey Guys,
I am setting up a new dedicated Postgres server, and will serve about
60 databases to a web site serving 250,000 people at the rate of about
20,000 a day. That may all be irrellevent though for the purposes of
this conversation.
The main thing about the application is that we're
Hello,
I try INITIALLY DEFERRED clause. In doc is relation between general
constraint and this clause. But combination with CHECK clause ends with
err message misplaced INITIALLY DEFERRED clause.
Is possible use INITIALLY DEFERRED with other variants constraint than
REFERENCES?
Thank You
As the documentation leads me to expect, my Postgresql 7.4
installation produces:
select substring('Bar, Foo' FROM '.') - 'B'
but even though my regex_flavor = advanced,
select substring('Bar, Foo' FROM '\\w') - NULL
select substring('Bar, Foo' FROM '***:\\w') - NULL
whereas I
On Wed, 2004-12-22 at 09:15, John Cunningham wrote:
Hey Guys,
I am setting up a new dedicated Postgres server, and will serve about
60 databases to a web site serving 250,000 people at the rate of about
20,000 a day. That may all be irrellevent though for the purposes of
this conversation.
To everyone who put in what's clearly a lot of time in building the new
website, well done! I think the combination of the 8.0 release and a
dramatically more professional website will do wonders for PostgreSQL.
Kudos!
---(end of broadcast)---
TIP
On Wed, Dec 22, 2004 at 01:32:14PM +, vinita bansal wrote:
I want to turn off the default setting in postgres for index creation on
primary key of a table. Is it possible and how?
PostgreSQL uses the index to enforce the primary key's uniqueness.
What problem are you trying to solve by
Hi all
Can anybody please help me in this?
Regards
Srini
--- srini vasan [EMAIL PROTECTED] wrote:
Hi
I am facing some issues with select query. The
values of the columns in one table contains \n as
a
part of the value. So when I execute the select
query
on this table, I am
Hi,
I want to turn off the default setting in postgres for index
creation on primary key of a table. Is it possible and how?
That is not possible, because the index is used to guarantee
the uniqueness of the primary key.
What is the reason you want to turn it off?
Sander.
John Cunningham wrote:
...
The machine in question will do nothing but serve databases. It's a
dual 3.2Ghz Xeon with 100GB or 15K RPM RAID 5 and 8 GB of RAM. I'd
like to configure it to get the most out of the server possible as far
as shared memory, sort memore, etc. I haven't found a lot of
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Dec 22, 2004, at 10:15 AM, John Cunningham wrote:
like to configure it to get the most out of the server possible as far
as shared memory, sort memore, etc. I haven't found a lot of
documentation on this.
Michael Fuhr [EMAIL PROTECTED] writes:
Was there a change affecting client certificate handling a couple
of months ago?
Yes, but it was pre-beta3.
However, this being on Windows ... I don't think the SSL code was
enabled at all in the Windows port as of beta3. I find this post-beta3
CVS log
On Wed, Dec 22, 2004 at 07:50:56AM -0800, srini vasan wrote:
Hi,
Can anybody please help me in this?
You could use the replace(text,text,text) function:
alvherre=# create table foo (a text);
CREATE TABLE
alvherre=# insert into foo values ('hola
alvherre'# ');
INSERT 155224 1
alvherre=#
sleect function(ssd_a) from ...
where function is one of the built-in text function or one of your own written
in some of the supported (by your version) procedure languages.
Regards.
En un mensaje anterior, srini vasan escribió:
--- srini vasan [EMAIL PROTECTED] wrote:
Hi
I am
On Tue, Dec 21, 2004 at 10:54:27AM -0500, Tom Lane wrote:
Patrick Hatcher [EMAIL PROTECTED] writes:
Curious: Why would a DELETE FROM tablename work while a pg_dump is
occurring but a TRUNCATE tablename will stay in a lock state until the
pg_dump is complete?
TRUNCATE requires an
On Wed, Dec 22, 2004 at 04:17:29PM +0100, Pavel Stehule wrote:
I try INITIALLY DEFERRED clause. In doc is relation between general
constraint and this clause. But combination with CHECK clause ends with
err message misplaced INITIALLY DEFERRED clause.
The CREATE TABLE documentation for
I'm running Red Hat Enterprise ES with all the most recent updates.
The error - in initdb - was that the system couldn't find
ascii_and_mic libraries. 7.3.6 ran without a hitch.
The RAID 1+0 - is that a stripped / mirrored condifuration? How big
of a difference will that make in performance do
I like the idea generally, just some comments:
On Tue, Dec 21, 2004 at 07:09:39PM -0500, Paul Tillotson wrote:
(1) VACUUM KEEP_EARLY_FREE_PAGES mybloatedtable; -- item (a)
This may not be necessary anyway, but could improve performance.
(2) UPDATE mybloatedtable SET foo = foo WHERE ctid
Ned Lilly wrote:
To everyone who put in what's clearly a lot of time in building the new
website, well done! I think the combination of the 8.0 release and a
dramatically more professional website will do wonders for PostgreSQL.
Kudos!
Amen.
Mike Mascari
---(end of
vinita bansal [EMAIL PROTECTED] writes:
I want to turn off the default setting in postgres for index creation on
primary key of a table. Is it possible and how?
No. The index is needed to enforce the unique constraint.
regards, tom lane
Yes, all versions were installed through the win installer.
I didn't change the SSL settings from default, so I believe it should not be
using it. In my haste, I did not keep a copy of the logs generated during
the RC1 installation. I can try RC2 this evening (outside of business
hours) and let
=?iso-8859-2?Q?Egy=FCd_Csaba?= [EMAIL PROTECTED] writes:
# explain analyze select meterid, tstamp, pp, pm, status from measured_1
where tstamp = '2004.12.22 00:00' and tstamp = '2004.12.22 23:59' order by
tstamp, meterid;
Sort (cost=2619.02..2622.78 rows=1505 width=42) (actual
Brandon Craig Rhodes [EMAIL PROTECTED] writes:
As the documentation leads me to expect, my Postgresql 7.4
installation produces:
select substring('Bar, Foo' FROM '.') - 'B'
but even though my regex_flavor = advanced,
select substring('Bar, Foo' FROM '\\w') - NULL
select
Pavel Stehule [EMAIL PROTECTED] writes:
Is possible use INITIALLY DEFERRED with other variants constraint than
REFERENCES?
No, we only support deferring foreign key constraints at the moment.
regards, tom lane
---(end of
On Wed, 22 Dec 2004 10:38:01 -0600, John Cunningham [EMAIL PROTECTED] wrote:
I'm running Red Hat Enterprise ES with all the most recent updates.
The error - in initdb - was that the system couldn't find
ascii_and_mic libraries. 7.3.6 ran without a hitch.
That's very odd, cause I'm using
On Wed, Dec 22, 2004 at 03:09:08PM +0100, Együd Csaba wrote:
Hi,
Is it a normal behavior that if I give a where clause with an existent index
key, then postgres uses the index, but if I give it a non existent value
than it refuses to use the index.
Whether the value exists is irrelevent. What
Hi,
I am actually migrating indexes from oracle database to postgres. I wanted
to turn it off so that index on the same columns is not created again (index
created for primary key of a table). I'll probably need to check in that
case and not create the index if it is on the primary key of the
MF == Michael Fuhr [EMAIL PROTECTED] writes:
MF On Tue, Dec 14, 2004 at 06:34:05PM +0100, Christian Kratzer wrote:
At least on FreeBSD 5.x there is no need to build a customer kernel.
The following can be set in /boot/loader.conf
MF I forgot about /boot/loader.conf. If I get a chance I'll
a == alex [EMAIL PROTECTED] writes:
a We are currently looking at Dell / HP
a but the questions is
a - how many processors (2 or 4)
a - do we gain with 4 cpus if we probably never have a few users connected
a - what processors are recommended Opteron / Xeon / Itanium
a - how much memory ?
Hi Tom,
may be it was misunderstandable...
I meant that there is no rows with tstamp='2004.12.22 00:00'. Certainly
there are rows matching the range (00:00 - 23:59); from 13:00 to 23:00, but
not before 13:00.
The only difference betwen the two queries is this value. If the begining of
the
The new site launched, and the download url has changed, it is now:
http://www.postgresql.org/download/mirrors-ftp
(or maybe http://wwwmaster.postgresql.org/download/mirrors-ftp, but I'm
not sure. They both work for now.)
On Tue, 2004-12-21 at 21:37, Marc G. Fournier wrote:
Its been 3 weeks
On Wed, 22 Dec 2004, Scott Marlowe wrote:
The new site launched, and the download url has changed, it is now:
http://www.postgresql.org/download/mirrors-ftp
I don't see my mirror (Russia) ftp.ru.postgresql.org !
(or maybe http://wwwmaster.postgresql.org/download/mirrors-ftp, but I'm
not sure.
OK Guys - here's the config file as I've writtten it.
I'll paste in the whole thing before, but this is the important stuff:
max_connections = 256
shared_buffers = 32768 # (256 MB)
sort_mem = 1024 # min 64, size in KB
fsync = No
wal_sync_method = fsync # the default varies across platforms:
On Wed, Dec 22, 2004 at 17:09:26 +,
vinita bansal [EMAIL PROTECTED] wrote:
I am actually migrating indexes from oracle database to postgres. I wanted
to turn it off so that index on the same columns is not created again
(index created for primary key of a table). I'll probably need to
=?iso-8859-2?Q?Egy=FCd_Csaba_=28Freemail=29?= [EMAIL PROTECTED] writes:
The point is that there are cases where a primary key index is not used -
even if the condition is formaly good.
You haven't actually shown us such a case. In the case you gave,
I think the planner probably made the right
On Wed, 22 Dec 2004 12:28:22 -0500, Vivek Khera khera@kcilink.com wrote:
a == alex [EMAIL PROTECTED] writes:
a We are currently looking at Dell / HP
a but the questions is
a - how many processors (2 or 4)
a - do we gain with 4 cpus if we probably never have a few users connected
a -
But why? I thought the planner is for choose the quicker way to the target
point. If there is an index which is probably would speed up the query then
why does the planner ignore that?
The difference between the result times is 16x. I can't understand why the
planner thinks it is the better
A. Mous [EMAIL PROTECTED] writes:
I didn't change the SSL settings from default, so I believe it should not be
using it.
Nope, the default libpq behavior is prefer SSL, so it will try an SSL
connection first ... if it is compiled to support SSL, which I believe
was not true in the beta3 windows
Ok, then the problem has nothing to do with SSL. IIRC, you get this
error message if you have SSL support compiled in (which the MSI version
does) and get a failure really early (typical example: connection is
accepted and then dropped right away). It has nothing to do with SSL.
The output of
On Wed, Dec 22, 2004 at 18:44:00 +0100,
Együd Csaba (Freemail) [EMAIL PROTECTED] wrote:
But why? I thought the planner is for choose the quicker way to the target
point. If there is an index which is probably would speed up the query then
why does the planner ignore that?
Because doing an
Thanks. I'll have to wait until after business hours to try RC2 and play a
bit more. Right now it's happily running Beta3. I'll be more diligent in
collecting log data with my next attempt.
Much thanks,
-Peter
-Original Message-
From: Magnus Hagander [mailto:[EMAIL PROTECTED]
Sent:
=?iso-8859-2?Q?Egy=FCd_Csaba_=28Freemail=29?= [EMAIL PROTECTED] writes:
The difference between the result times is 16x. I can't understand why the
planner thinks it is the better way...
The planner thinks that because it thinks the second query will fetch
1500 times as many rows as the first.
On Wed, 2004-12-22 at 11:30, John Cunningham wrote:
OK Guys - here's the config file as I've writtten it.
I'll paste in the whole thing before, but this is the important stuff:
max_connections = 256
Are you using a connection pooling scheme (jdbc based pooling, pgpool,
etc...)? If not,
On Wed, 2004-12-22 at 11:41, Lonni J Friedman wrote:
On Wed, 22 Dec 2004 12:28:22 -0500, Vivek Khera khera@kcilink.com wrote:
a == alex [EMAIL PROTECTED] writes:
a We are currently looking at Dell / HP
a but the questions is
a - how many processors (2 or 4)
a - do we gain with 4
On Dec 22, 2004, at 12:41 PM, Lonni J Friedman wrote:
get expected performance from those boxes. They seem to do
something to the RAID controllers to make them not work as fast as one
would expect the equivalent name-brand part (eg, LSI RAID card or
Adaptec RAID card) and similar disk drives.
On Dec 22, 2004, at 1:09 PM, Scott Marlowe wrote:
I've use the Dell PERC 4DC and had VERY good performance from it. IT's
the late model U320 LSI MegaRAID and runs great. I do remember that
the
2650 and few other Dells had the serverworks chipset in them that
caused
a lot of context switches in
Yes, I vacuum analyze it once in every hour.
Thank you for the suggestions, now I'm getting understand the point...
With default_statistics_target=500 and random_page_cost=0.1 and reanalyzing
the db the planner always uses the index. Certainly if there were more rows
in the table the planner
On Dec 22, 2004, at 12:09 PM, vinita bansal wrote:
I am still not clear on why postgres has this restriction?
By uniqueness, you mean to say that if later anyone wants to add a primary key constraint on a table which already has a primary key defined, postgres will use this index to determine
Hi,
I am actually migrating indexes from oracle database to postgres. I wanted
to turn it off so that index on the same columns is not created again
(index created for primary key of a table). I'll probably need to check in
that case and not create the index if it is on the primary key of the
The server is a DELL Poweredge 2650 with it's built in RAID - 4 disks
currently in a RAID 5 config. I will check on the battery backup.
I'm putting this server together and rebuilding our overall db
structure all at the same time, so I have a good amount of flexiblity.
I realized I was not
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Dec 22, 2004, at 2:36 PM, John Cunningham wrote:
The shared buffers was a big concern - I've read that there's a limit
that helps, but as the machine will only do DB transactions, I don't
know what else to do with the RAM. It's intended for PG's
I attempted to install 8.0 RC 2 alongside 7.4.5 on my OS X box, but
initdb failed with an error about not enough shared memory.
Remembering that this was a problem for starting two postmasters at the
same time on OS X, I increased the shmmax value to 500 megabytes (I had
seen something say
John Cunningham [EMAIL PROTECTED] writes:
The shared buffers was a big concern - I've read that there's a limit
that helps, but as the machine will only do DB transactions, I don't
know what else to do with the RAM. It's intended for PG's use.
The kernel will use it for disk caching which
Jeffrey Melloy [EMAIL PROTECTED] writes:
I attempted to install 8.0 RC 2 alongside 7.4.5 on my OS X box, but
initdb failed with an error about not enough shared memory.
Don't forget that both shmmax and shmall may need attention ... and,
just to confuse matters, they are measured in different
What version of OS X?
Apparently some of the earlier versions did not permit changing this parameter without recompiling the kernel. It seems to have been changed in the more recent versions, though:
http://www.opendarwin.org/pipermail/hackers/2002-August/003583.html
Tom Lane wrote:
Jeffrey Melloy [EMAIL PROTECTED] writes:
I attempted to install 8.0 RC 2 alongside 7.4.5 on my OS X box, but
initdb failed with an error about not enough shared memory.
Don't forget that both shmmax and shmall may need attention ... and,
just to confuse matters, they are
Can anyone do something about the duplicate messages from the list?
It seems to be delivering twice from svr1.postgresql.org to
commandprompt.com (the timestamps on the received headers are the same
up until that point). Delivering to hosting.commandprompt.com doesnt
seem to duplicate. Nor does
I need to convert an existing unique index on a very heavily inserted table
into a primary key. Alter table works, but locks the table for too long.
As a hack, can I just set pg_index.indisprimary = 't' and
pg_constraint.contype = 'p' for the appropriate row in each? This is for
7.4.6 and
Hi everyone,
I am new to postgresql.
I am planning to migrate our current database (mysql) to posgresql. I
am currently having difficulties dealing with the order of the tables
that is created.
Here is what I do:
1. I use myqsldump to get the schema and data.
2. I use my2pg and a java utility I
1. Create a script that creates the tables without RI stuff like
foreign keys and triggers.
2. Import the data into the tables using COPY
3. Create the RI stuff with 'alter table' and 'create trigger'
statements
http://www.postgresql.org/docs/current/static/sql-copy.html
Ed L. [EMAIL PROTECTED] writes:
I need to convert an existing unique index on a very heavily inserted table
into a primary key. Alter table works, but locks the table for too long.
As a hack, can I just set pg_index.indisprimary = 't' and
pg_constraint.contype = 'p' for the appropriate
Hi all,
I've installed
TSearch2 with some success- my table now contains a tsvector field that's
indexed and I can run full text queries.
However, the trigger
that I created to update the full text index when a row is modified appears to
cause a problem.It's returning the error when I run
Hi,
I have a stored proc which is for filling 2 tables with empty rows in every
minutes. If the server has been stopped for more then 3 hours the insertion
takes too much.
I switched off the insert execution, and debugged the proc and realized that
the loop increasing the timestamp takes so long.
Title: bytea
We have column job_data with datatype bytea. This is one of the columns in table used by quartz scheduler(open source). We have to insert the data into it.
The data is as follows.
72 matches
Mail list logo