Thread: refcurosr vs. setof

Hello,

I need to write several PL/pgSQL functions all returning a "result set" wich
can be obtained by a single SELECT statement.
For now the functions are called by a Java application.
Both REFCURSOR and SETOF serve my purpose, but I was wondering if there is a
perfonance difference between the two. The result set can become quite
large.

I hope not to ask this question the 1001 time, though I couldn't find
anything on the net.. Any hints are welcome.

Regards
R�diger

On Sun, Apr 17, 2005 at 10:05:29PM +0200, Rüdiger Herrmann wrote:
>
> I need to write several PL/pgSQL functions all returning a "result set" wich
> can be obtained by a single SELECT statement.
> For now the functions are called by a Java application.
> Both REFCURSOR and SETOF serve my purpose, but I was wondering if there is a
> perfonance difference between the two. The result set can become quite
> large.

Here's an excerpt from the "Control Structures" section of the
PL/pgSQL documentation:

    The current implementation of RETURN NEXT for PL/pgSQL stores
    the entire result set before returning from the function, as
    discussed above.  That means that if a PL/pgSQL function produces
    a very large result set, performance may be poor: data will be
    written to disk to avoid memory exhaustion, but the function
    itself will not return until the entire result set has been
    generated....Currently, the point at which data begins being
    written to disk is controlled by the work_mem configuration
    variable.

You might want to test both ways in typical and worst-case scenarios
and see how each performs.

--
Michael Fuhr
http://www.fuhr.org/~mfuhr/