ivan marchesini napisał(a):
Dear postgres Users,
I have a simple question I think.
I have a table that contains some columns
one of these columns (the columns ID) contains distinct integer
values ...
I need to insert into this table some other records but I only need that
values were inserted int
On Fri, 11 May 2007 19:09:07 +0500
imad <[EMAIL PROTECTED]> wrote:
> You might be looking for PostgreSQL RECORD data type.
Thanks. Give me an example please. I saw the documentation already.
>
> --Imad
> www.EnterpriseDB.com
>
> On 5/11/07, Gábriel Ákos <[EMAIL PROTECTED]> wrote:
> > Hi,
> >
>
Hi,
How should I define a record type (there is no table with this record
type) programmatically in pl/pgsql?
I'd like to return a record with 3 string elements, 2 integers and 1
date.
Rgds,
Akos
--
Üdvözlettel,
Gábriel Ákos
-=E-Mail :[EMAIL PROTECTED]|Web: http://www.i-logic.hu =-
-=Tel/fax
You might be looking for PostgreSQL RECORD data type.
--Imad
www.EnterpriseDB.com
On 5/11/07, Gábriel Ákos <[EMAIL PROTECTED]> wrote:
Hi,
How should I define a record type (there is no table with this record
type) programmatically in pl/pgsql?
I'd like to return a record with 3 string elements
You can use CREATE TYPE:
http://www.postgresql.org/docs/8.2/interactive/sql-createtype.html
Example from the documentation:
CREATE TYPE compfoo AS (f1 int, f2 text);
Then make your function return compfoo (or setof compfoo).
Alternately, you can define your function with out or in/out
par
I have a need to convert an incorrectly typed serial column to a
regular integer column. Basically this just involves removing the
sequence. I am able to successfully remove the default value (DROP
DEFAULT) (which seems to use nextval) and now pgadmin does show the
column as an integer, but I ca
On 5/11/07, Collin Peters <[EMAIL PROTECTED]> wrote:
I have a need to convert an incorrectly typed serial column to a
regular integer column. Basically this just involves removing the
sequence. I am able to successfully remove the default value (DROP
DEFAULT) (which seems to use nextval) and no
create a function with return type as a RECORD.
CREATE FUNCTION xyz() RETURNS record AS
$$
declare
abc RECORD;
begin
abc := (1, 2);
return abc;
end;
$$
language plpgsql;
And execute the function in this fashion:
select a, b from xyz() as (a int, b int);
Do you like that ... :-)
--Imad
www
Hi Gabriel,
There are two ways to do this:
1. Imad's way (Define the function with the return type as RECORD). Its only
problem is that while querying from this function, you need to give a proper
SELECT query or else PG returns an error.
e.g. As Imad gives in his example ...
CREATE FUNCTION x
On 5/11/07, Robins <[EMAIL PROTECTED]> wrote:
Hi Gabriel,
There are two ways to do this:
1. Imad's way (Define the function with the return type as RECORD). Its only
problem is that while querying from this function, you need to give a proper
SELECT query or else PG returns an error.
Yeah ...
"=?ISO-8859-1?Q?Rodrigo_De_Le=F3n?=" <[EMAIL PROTECTED]> writes:
> On 5/11/07, Collin Peters <[EMAIL PROTECTED]> wrote:
>> Is there any way to remove the sequence fully?
> ALTER SEQUENCE dtab_i_seq OWNED BY NONE;
Pre-8.2 that command doesn't exist, but you can get the same effect if
you manually
Anything pre-8.2?
On 5/11/07, Rodrigo De León <[EMAIL PROTECTED]> wrote:
CREATE TABLE dtab (i SERIAL);
ALTER TABLE dtab ALTER COLUMN i DROP DEFAULT;
ALTER SEQUENCE dtab_i_seq OWNED BY NONE;
DROP SEQUENCE dtab_i_seq;
---(end of broadcast)---
Just to be sure I am getting this right:
I have a big table I want to partition:
create table big_table (
row_date timestamp with time zone,
row_data character varying(80)
};
A nice solution would be to spread its rows in one of 12 child tables
according to which month the date f
13 matches
Mail list logo
