On Tue, May 31, 2011 at 5:01 PM, Tom Lane wrote:
> Merlin Moncure writes:
>> I've never taken the time to really get my head around 'lateral'
>> enough to say for sure if it provides clean workarounds for all the
>> cases that get people into hot water. The case that used to get me a
>> lot is (
Merlin Moncure writes:
> I've never taken the time to really get my head around 'lateral'
> enough to say for sure if it provides clean workarounds for all the
> cases that get people into hot water. The case that used to get me a
> lot is (the unfortunately generally under utilized) custom aggre
On Tue, May 31, 2011 at 3:28 PM, Tom Lane wrote:
> Merlin Moncure writes:
>> There have been multiple complaints about this in the archives. In
>> the old days, you would have to rewrite your query to use the 'select
>> * from func()' form (which isn't always so easy) or use a subquery and
>> th
Merlin Moncure writes:
> There have been multiple complaints about this in the archives. In
> the old days, you would have to rewrite your query to use the 'select
> * from func()' form (which isn't always so easy) or use a subquery and
> the 'offset 0' hack. Running in to this problem has actua
On Tue, May 31, 2011 at 11:57 AM, David Johnston wrote:
> See my thoughts below. Other user's opinions (or a pointer to where this
> topic has been previously discussed) are greatly welcomed.
> Thank you for the technical detail on how ().* gets expanded by the engine.
> I still believe it would
Cc: Tom Lane; pgsql-general@postgresql.org
> Subject: Re: [GENERAL] Function Column Expansion Causes Inserts To Fail
>
> The basic issue is that:
> select (func()).*, if the return type has fields, 'a', 'b', gets expanded
to:
>
> select (func()).a, (func()).b;
On Tue, May 31, 2011 at 9:24 AM, David Johnston wrote:
> From syntax works fine for literals but how would you then get table.column
> values into the function call - where you want to evaluate multiple rows
> from the source table? In order to feed rows to a function you need the
> function in t
> -Original Message-
> From: pgsql-general-ow...@postgresql.org [mailto:pgsql-general-
> ow...@postgresql.org] On Behalf Of Tom Lane
> Sent: Monday, May 30, 2011 11:10 PM
> To: David Johnston
> Cc: pgsql-general@postgresql.org
> Subject: Re: [GENERAL] Function Co
"David Johnston" writes:
> SELECT ( createpkrecord('1')).*;
> [ results in function being called more than once ]
Yeah. Don't do that. Better style is
SELECT * FROM createpkrecord('1');
regards, tom lane
--
Sent via pgsql-general mailing list (pgsql-general@pos
PostgreSQL 9.0.4
The following script fails even though the "pkonlytest" table is empty since
we just created it.
>> BEGIN SCRIPT
CREATE TABLE pkonlytest (
pkid text PRIMARY KEY
);
CREATE OR REPLACE FUNCTION createpkrecord(INOUT pkvalue text, OUT col
10 matches
Mail list logo