On Mon, Jul 11, 2016 at 12:06 AM, Tim Dawborn wrote:
> tmp=# INSERT INTO foo (a, b, c, d) VALUES (1, 2, 'four', true)
> tmp-# ON CONFLICT (a, b) WHERE d = true
> tmp-# DO UPDATE SET c = 'four' WHERE foo.a = 1 AND foo.b = 2 AND foo.d =
> true;
> ERROR: there is no unique or exclusion constraint ma
Hi all,
I'm struggling to work out the correct way to use a partial unique index in
an upsert context.
Here's the context: I'm creating a partial unique index on a table:
tmp=# CREATE TABLE foo (a INT NOT NULL, b int NOT NULL, c TEXT, d BOOLEAN
DEFAULT false);
CREATE TABLE
tmp=# CREATE UNIQUE IN
On 07/11/2016 01:00 AM, arnaud gaboury wrote:
On Sun, Jul 10, 2016 at 7:07 PM, Adrian Klaver
wrote:
On 07/10/2016 03:13 AM, arnaud gaboury wrote:
My box upgraded postgresql from 9.4.8 to 9.5.3 when doing a full
version upgrade (Fedora 23 --> 24).
I have now :
/db/pgsal/data > initialize
Adrian Klaver wrote:
> There is still the issue of the 9.4 binaries and which version of
> Postgres you are running commands from at any given time. I would
> suggest using full paths to the commands until this is resolved.
It reads to me as if the OP's Fedora 23 -> Fedora 24 upgrade didn't
instal
On Fri, Jul 1, 2016 at 05:39:28PM -0400, Bruce Momjian wrote:
> > Replacing beta1 with beta2 leads to a 404 (for both direct download links)
>
> Yes, I confirmed in a later email that it is still broken.
>
> It was also reported by someone else today:
>
>
> https://www.postgresql.org/mes
On Mon, Jul 11, 2016 at 8:45 AM, Deven Phillips
wrote:
> We need to update a JSONB data structure in our tables to include an
> 'is_valid' flag. I was able to build a CTE which I use to create a temp
> table containing the 'is_valid' value, the path where it needs to be set,
> and the join criter
We need to update a JSONB data structure in our tables to include an
'is_valid' flag. I was able to build a CTE which I use to create a temp
table containing the 'is_valid' value, the path where it needs to be set,
and the join criteria. Let's say that the temp table looks like:
id TEXT,
> time_in
>> On 2016-07-09 08:20, haman...@t-online.de wrote:
>> > Hi,
>> >
>> > a table is associated with a trigger for normal use.
>> > An admin (someone with ALTER privilege) can disable tthe trigger, run some
>> > bulk update,
>> > and then re-enable it. This means, however, that normal user activity
>> On 2016-07-09 08:20, haman...@t-online.de wrote:
>> > Hi,
>> >
>> > a table is associated with a trigger for normal use.
>> > An admin (someone with ALTER privilege) can disable tthe trigger, run some
>> > bulk update,
>> > and then re-enable it. This means, however, that normal user activity
Hi Adi: ( Is this correct? Adi is what your message uses to mark your replies ).
On Sun, Jul 10, 2016 at 6:53 PM, Prashanth Adiyodi
wrote:
> Hi Franciso, My comments below inline
Got them. Only problem is your MUA does not signal quotes. It looks
like some kind of outlook by the headers it sends
Hi
> On Jul 08, John McKown modulated:
> ...
>> I think the "problem" that he is having is fixable only by changing how
>> PostgreSQL itself works. His problem is a PL/pgSQL function which is
>> 11K lines in length. When invoked, this function is "compiled" into a
>> large tokenized parse tree. Th
On Sun, Jul 10, 2016 at 7:07 PM, Adrian Klaver
wrote:
> On 07/10/2016 03:13 AM, arnaud gaboury wrote:
>>
>> My box upgraded postgresql from 9.4.8 to 9.5.3 when doing a full
>> version upgrade (Fedora 23 --> 24).
>>
>> I have now :
>>
>> /db/pgsal/data > initialized with 9.5, with postgresql.co
12 matches
Mail list logo
