Richard Broersma writes:
> Actually I'm still confused. I must me missing something. When I
> manually following the directions of:
> http://www.postgresql.org/docs/8.4/interactive/queries-with.html
> I get the following when I try:
> WITH RECURSIVE t(n) AS (
> VALUES (1)
> UNION ALL
>
On Wed, Nov 4, 2009 at 2:11 PM, Tom Lane wrote:
> Richard Broersma writes:
>> Can anyone one explain why a "WITH RECURSIVE" query has the same
>> results regardless whether UNION or UNION ALL is specified?
>
> Well, if the rows are all different anyway, UNION isn't going to
> eliminate any ...
On Wed, Nov 4, 2009 at 2:11 PM, Tom Lane wrote:
> Richard Broersma writes:
>> Can anyone one explain why a "WITH RECURSIVE" query has the same
>> results regardless whether UNION or UNION ALL is specified?
>
> Well, if the rows are all different anyway, UNION isn't going to
> eliminate any ...
O
Richard Broersma writes:
> Can anyone one explain why a "WITH RECURSIVE" query has the same
> results regardless whether UNION or UNION ALL is specified?
Well, if the rows are all different anyway, UNION isn't going to
eliminate any ...
regards, tom lane
--
Sent via pgs
Can anyone one explain why a "WITH RECURSIVE" query has the same
results regardless whether UNION or UNION ALL is specified?
broersr=> WITH RECURSIVE t(n) AS (
broersr(> VALUES (1)
broersr(> UNION
broersr(> SELECT n+1 FROM t WHERE n < 100
broersr(> )
broersr-> SELECT sum(n) FROM t;
sum
I've just tracked down a serious bug with PG_BYTEA columns, which is probably
perl specific. But I thought
people might want to know:
Package: libdbd-pg-perl
Version: upgrading to 2.8.7 compared to 1.49. Our 1.49 was patched for the
memory leak in BYTEA.
With 2.8.7, if pg_server_prepare is s
On Wed, Nov 4, 2009 at 1:14 PM, Leif Biberg Kristensen
wrote:
> On Wednesday 4. November 2009 21.03.26 Scott Marlowe wrote:
>> On Wed, Nov 4, 2009 at 11:53 AM, Leif Biberg Kristensen
>> > This looks strange to me, but it works:
>> >
>> > pgslekt=> CREATE TABLE participant_notes (
>> > pgslekt(>
On Wednesday 4. November 2009 21.03.26 Scott Marlowe wrote:
> On Wed, Nov 4, 2009 at 11:53 AM, Leif Biberg Kristensen
> > This looks strange to me, but it works:
> >
> > pgslekt=> CREATE TABLE participant_notes (
> > pgslekt(> person_fk INTEGER NOT NULL,
> > pgslekt(> event_fkINTEGER
On Wed, Nov 4, 2009 at 11:53 AM, Leif Biberg Kristensen
wrote:
> On Wednesday 4. November 2009 19.37.41 Scott Marlowe wrote:
>> On Wed, Nov 4, 2009 at 11:36 AM, Leif Biberg Kristensen
>> wrote:
>>> I'd missed that particular syntax.
>>>
>>> This table is now without a primary key, but is that a p
On Wednesday 4. November 2009 19.37.41 Scott Marlowe wrote:
> On Wed, Nov 4, 2009 at 11:36 AM, Leif Biberg Kristensen
> wrote:
>> I'd missed that particular syntax.
>>
>> This table is now without a primary key, but is that a problem? I don't
>> expect it to grow beyond maybe a few thousand rows.
On Wed, Nov 4, 2009 at 11:36 AM, Leif Biberg Kristensen
wrote:
> On Wednesday 4. November 2009 19.24.29 Scott Marlowe wrote:
>> You're referencing a single column, which does not have a unique key
>> on it. Being part of a two column unique PK index doesn't count, as
>> you could have an entry wh
On Wednesday 4. November 2009 19.24.29 Scott Marlowe wrote:
> You're referencing a single column, which does not have a unique key
> on it. Being part of a two column unique PK index doesn't count, as
> you could have an entry where one column or the other repeats on its
> own while the other colu
On Wed, Nov 4, 2009 at 11:03 AM, Leif Biberg Kristensen
wrote:
> PostgreSQL 8.3.8 on Gentoo Linux.
>
> I've got a junction table:
>
> CREATE TABLE participants (
> person_fk INTEGER REFERENCES persons (person_id),
> event_fk INTEGER REFERENCES events (event_id) ON DELETE CASCADE,
> sort_o
PostgreSQL 8.3.8 on Gentoo Linux.
I've got a junction table:
CREATE TABLE participants (
person_fk INTEGER REFERENCES persons (person_id),
event_fk INTEGER REFERENCES events (event_id) ON DELETE CASCADE,
sort_order INTEGER NOT NULL DEFAULT 1,
is_principal BOOLEAN NOT NULL DEFAULT
14 matches
Mail list logo