> These pointers are useless
> except in the very unusual case where one steps forward and then back
> in a sequential scan (for example, "FETCH 1; FETCH BACKWARD 1;" in a
> cursor).
Actually I think it should be known whether a fetch backward is actually
allowed for a particular cursor, since it
Hi,
I am trying to use postgresql in some research
project. I need to add some new operators, some
will probably take more than 2 input tuple streams.
The new operator will be considered by optimizer
when estimating cost and choosing a plan.
Can anyone tell me how hard this will be? And
where
Hi all,
I've posted this message twice nefore on pgsql-bugs, but I can't seem to get
any response. The issue described below is pretty serious (at least as far as
my company is concerned):
While trying to upgrade Postgres from version 7.0.3 to 7.1.2, I find that a
program which dlopens the li
Compare price of implementation.
For that $100k on the oracle license you can toss in a few more gigs
of memory and a few extra CPU's and perhaps 15k drives rather than 10k
ones :)
Then toss in the monthly support contracts between Oracle & Great
Bridge (or Pgsql.inc if you can get anyone on the
I tried another group without success, so
The documents do not explain clearly enough, and here are no examples
showing how to connect (via TCP/IP) to pg on another host.
I need to supply user, password and database/
Here are examples that do not work:
// exec sql connect "tcp:postgresql:
"Jan Wieck" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> mlw wrote:
> > [...]
> > REPLACE into table set xx=yy, ww = zz where ID = fubar;
> >
> > A MUCH better solution!
>
> Please solve the trigger problem at least theoretical before
> claiming that mySQL is that MUC
Hi,
today I had a little problem with the actual
JDBC driver. Due to the use of Runtime.addShutdownHooks(), which throws a
SecurityException if used within a servlet container (WebLogic Server 5.1 &
6.0), the postgres JDBC driver is unusable in this environments. I spent some
minutes looki
If one runs Postgresql with fsync turned off in order to get better
performance, is there a risk that the database internals will be
damaged by a system crash? I realize that database discrepancies are
possible, but I am referring to a failure where storage allocation
gets messed up, or critical i
Frequently one wants a data set returned in the same order as the
index used in the query. Informix (at least) has implicit order-by,
which means that the data will be returned in collating order if the
query forces use of the appropriate index.
Does Postgresql do this?
If not, does an Order-by
Dear Tom,
>
>
> [EMAIL PROTECTED] (Robert E. Bruccoleri) writes:
> > For my immediate problem, would removing the spinlock acquisition
> > be OK?
>
> It'd be interesting to remove the marked lines:
>
> bufHdr = &BufferDescriptors[buffer - 1];
> - SpinAcquire(BufMgrLock);
>
[EMAIL PROTECTED] (Robert E. Bruccoleri) writes:
> BTW, given the high level of support that you provide to the PostgreSQL
> community, it's very accurate to state that support for PostgreSQL
> is far superior to that of Oracle, especially for SGI systems.
It's all about having the source code av
Bruce Momjian wrote:
>
> > Sending this to the Hackers list because I think that there may be a bug.
>
> Did you include the smell in the email message? :-)
>
What sort of card do I need to read, I mean smell this?
Does it help that this machine runs Win95, thereby already stinking
somewhat?
Dear Tom,
>
>
> [EMAIL PROTECTED] (Robert E. Bruccoleri) writes:
> > BTW, given the high level of support that you provide to the PostgreSQL
> > community, it's very accurate to state that support for PostgreSQL
> > is far superior to that of Oracle, especially for SGI systems.
>
> It's all abo
I am planning to deal with all of the following TODO items today:
* pg_database should have unique indexes on oid and on datname.
* pg_shadow should have unique indexes on usename and on usesysid.
* pg_am should have unique index on oid
* pg_opclass should have unique index on oid
* pg_amproc
On Fri, 8 Jun 2001, Mike Cianflone wrote:
>> I'm running into some strange behavior with foreign keys which are a
>> tuple of primary keys.
>>
>>
>> I have a parent table sector, and a child of that is cell_area table and
a
>> child of that is unit table.
>>
>> The cell_area table has a
We have been researching replication for several months now, and
I have some opinions to share to the community for feedback,
discussion, and/or participation. Our goal is to get a replication
solution for PostgreSQL that will meet most needs of users
and applications alike (mission impossible the
I am trying to create a user defined C function that will
be called within PL/pgSQL
Namely, I need a function that will create a new Large Object and copy
the data
of an existing Large Object into the new Large Object.
This is the way the function would be registered
CREATE FUNCTION copyoid(oi
On Mon, 11 Jun 2001 19:46:44 GMT, you wrote:
>We have been researching replication for several months now, and
>I have some opinions to share to the community for feedback,
>discussion, and/or participation. Our goal is to get a replication
>solution for PostgreSQL that will meet most needs of us
I have run into a small snag with adding indexes to pg_shadow: the first
attempt to run a non-bootstrap-mode backend fails with 'user "postgres"
does not exist'. The reason it fails is that user postgres is added
to pg_shadow by global.bki, which is run after template1.bki, which is
where all the
Hi
I wonder why all large objects is placed in one system table
pg_largeobject ).
I just want to trow an idea, why not create a similiar table and added pglo
( something like pglo_tablename ).
This way, the large objects will be spread in every table that has the blob
type.
Thank's
Andy
Vincent Roberts <[EMAIL PROTECTED]> writes:
> I notice that the lo_create and lo_open functions require a postgres
> connection (PGConn)
Those are the client-side functions; naturally they need a connection
to the backend. For the server-side functions, look in
src/backend/libpq/be-fsstubs.c. F
"Andy Samuel" <[EMAIL PROTECTED]> writes:
> I wonder why all large objects is placed in one system table
> pg_largeobject ).
Why not? There's no tight limit on the size of a table.
> I just want to trow an idea, why not create a similiar table and added pglo
> ( something like pglo_tablename )
> Although
> Postgres-R is a synchronous approach, I believe it is the closest to
> the goal mentioned above. Here is an abstract of the advantages.
If you only want synchronous replication, why not simply use triggers ?
All you would then need is remote query access and two phase commit,
and
> From Andreas' comments, it seems that for his application he would like
> a different behavior, but frankly I'm not certain why the current
> behavior would be detrimental in the use case he mentioned. If SQL92
> requires that any query with "= NULL" be rejected as illegal
You don't mean me, n
> I had a baby girl on Tuesday. I am working through my
> backlogged emails
> today.
Congratulations -:)
Vadim
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
Alex Pilosov <[EMAIL PROTECTED]> writes:
> I noticed current wierd behaviour of a less/greater than comparisons of
> things involving inet/cidr:
> 10.1.2.3/8 is considered to be less than 10.0.0.0/32
And what's wrong with that? Essentially this comes from the conclusion
that 10/8 is less than 1
trust me ... girls are soo much fun ... *roll eyes* *watches for
lightening*
On Mon, 11 Jun 2001, Mikheev, Vadim wrote:
> > I had a baby girl on Tuesday. I am working through my
> > backlogged emails
> > today.
>
> Congratulations -:)
>
> Vadim
>
> ---(end of broad
Alex Pilosov <[EMAIL PROTECTED]> writes:
> What I have right now is rewriting a <<= b to use index plan :
> (a >= network(b)) && ( a <= broadcast(b) )
> However, that breaks down, since (for example)
> if a=10.1.2.3/32 and b = 10.1.2.0/24, broadcast(b) will be 10.1.2.255/24,
> but 10.1.2.255/24 i
Jim Mercer <[EMAIL PROTECTED]> writes:
> while you are in there, can you cahnge the print functions so that they
> are consistent?
I believe they are consistent in 7.1; leastwise, you will have to make
a pretty good argument why we should change them again. We had a very
long discussion that led
Alex Pilosov <[EMAIL PROTECTED]> writes:
> On Mon, 11 Jun 2001, Tom Lane wrote:
>> While there may not be a user-visible function for next-network-part,
>> that hardly matters since the special-indexqual stuff isn't user-visible
>> either.
> Well, since I'm making an indexqual clause, I do need a
mlw <[EMAIL PROTECTED]> writes:
> cdinfo=# create view test as select * from zsong ;
> ERROR: pg_atoi: error reading "2156109797": Numerical result out of
> range
Is this 7.1? I thought we'd fixed all the places that treated OID
values as signed.
regards, tom lane
---
> I thought we agreed to merge them already.
We did? Good. Will make it so.
> Also, could we modify
> IndexScanOK() to catch this case and do a heap scan?
How? Also, why?
regards, tom lane
---(end of broadcast)---
TIP
Interesting - my experience is that Access, at least, generally treats
NULL's correctly:
(This was done under Access 2000):
create table foo (name text(20))
insert into foo values ("bar");
insert into foo values ("bar");
insert into foo values ("bar");
insert into foo values ("bar");
insert into
33 matches
Mail list logo
