Re: Pipelined functions in Postgres

["Milen Kulev" <makulev(at)gmx(dot)net>]

Title: Nachricht

Talha,

do you know how much memory is consumed by the SETOF function ?

What happens with memory consumption of the function if

SELECT ename FROM emp WHERE sal > $1

 returns 10 mio rows ? 

I suppose that memory for the RECORD structure is immediately reused by the next record.

 

Regards, Milen 

  

-----Original Message-----
From: Talha Khan [mailto:talha(dot)amjad(at)gmail(dot)com]
Sent: Tuesday, September 19, 2006 11:08 PM
To: Milen Kulev
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: [PERFORM] Pipelined functions in Postgres

Hi Milen,

 

Pipelined function is a code that acts like a database table.

 

Inorder to use this functionality in postgres you would need to write the function like this

 

CREATE OR REPLACE FUNCTION get_test_data (numeric)
    RETURNS SETOF RECORD AS
$$
DECLARE
    temp_rec    RECORD;
BEGIN
    FOR temp_rec IN (SELECT ename FROM emp WHERE sal > $1)
    LOOP
        RETURN NEXT temp_rec;
    END LOOP;
    RETURN;
END;
$$ LANGUAGE plpgsql;
 

now inorder to call this function you would write the code as follows

 

SELECT * FROM get_test_data(1000) AS t1 (emp_name VARCHAR);
 

 

Regards

Talha Amjad

 

On 9/19/06, Milen Kulev <makulev(at)gmx(dot)net> wrote:

Hello Lister,
I am curios whether I can emulate the Oracle pipelined functions functionality in PG too (using RETURN NEXT ). For more
information and examples about Oracle pipelined functions see:
http://asktom.oracle.com/pls/ask/f?p=4950:8:8127757633768425921::NO::F4950_P8_DISPLAYID,F4950_P8_CRITERIA:4447489221109

I have used  pipeline functions in DWH enviromnent  with success and would like
To use similar concept in PG too.

Any help, examples , links and  shared experiences would be greately appreciated.

Best Regards.
Milen


---------------------------(end of broadcast)---------------------------
TIP 1: if posting/reading through Usenet, please send an appropriate
      subscribe-nomail command to majordomo(at)postgresql(dot)org so that your
      message can get through to the mailing list cleanly