Re: PERSISTANT PREPARE (another point of view) - Mailing list pgsql-sql

From Milan Oparnica
Subject Re: PERSISTANT PREPARE (another point of view)
Date
Msg-id g6iuol$202l$1@news.hub.org
Whole thread Raw
In response to Re: PERSISTANT PREPARE (another point of view)  ("Pavel Stehule" <pavel.stehule@gmail.com>)
Responses Re: PERSISTANT PREPARE (another point of view)
List pgsql-sql
Pavel Stehule wrote:
> Hello
> 
> this is near my implemenation of stored procedures - it's not far from
> your view on prepared statements. There  result of any unbound select
> is directly forwarded to client - there is little bit different
> syntax, but idea is same
> 
> 
> create procedure foo(params)
> as $$
> begin
>   select a, b, c from tab1 -- unbound select
>   select ....
> end;
> 
> and you can call with CALL statement. Curent implementation of SRF
> function in plpgsql sn't eficient for big results - it needs
> materialised table in memory. Pipeline Oracle's functions are better,
> but we are limited one thread architecture.
> 
> regards
> Pavel Stehule
> 

Yeah, this would be pretty the same.

I just didn't understand if this is already implemented in PG ?

I've tried to create a PROCEDURE in QueryTool of pgAdminIII and I get 
following result:

ERROR:  syntax error at or near "PROCEDURE"
LINE 1: CREATE PROCEDURE aatestubnd(refcursor, insklid integer, inda...               ^

********** Error **********

ERROR: syntax error at or near "PROCEDURE"
SQL state: 42601
Character: 8


If I try with CREATE FUNCTION I get this result (as expected):

ERROR:  function result type must be specified

********** Error **********

ERROR: function result type must be specified
SQL state: 42P13


Is this because of QueryTool's limitations or is this feature not yet 
implemented in Postgre ?

Though, CREATE PROCEDURE should be a valid SQL 92 standard procedure.

Best regards,

Milan Oparnica


pgsql-sql by date:

Previous
From: Milan Oparnica
Date:
Subject: Re: PERSISTANT PREPARE (another point of view)
Next
From: Jean-David Beyer
Date:
Subject: Re: Re: Efficiently determining the number of bits set in the contents of, a VARBIT field