partitioned table insert triggers

I'm trying to create a date-partitioned version of the postgres_log
table that is described here:
http://www.postgresql.org/docs/current/static/runtime-config-logging.html


In lieu of hardcoding the partition table names in the trigger, as shown
in the table partitioning docs, I wanted to dynamically compute the
table name. I tried:

        CREATE OR REPLACE FUNCTION postgres_log_insert_trigger()
        RETURNS TRIGGER AS $$
        DECLARE
            c text;
            ptable text :=3D 'postgres_log';
            ctable text :=3D ptable || '_' ||
        to_char(NEW.session_start_time, 'YYYY_MM');
        BEGIN
            c :=3D 'INSERT INTO ' || ctable || ' VALUES (NEW.*)';
            EXECUTE c;
            RETURN NULL;
        END;
        $$
        LANGUAGE plpgsql;

I get:
        ERROR:  NEW used in query that is not in a rule
        CONTEXT:  SQL statement "INSERT INTO
        measurement_postgres_log_2008_05 VALUES (NEW.*)"
        PL/pgSQL function "postgres_log_insert_trigger" line 7 at
        EXECUTE statement

I didn't really think that would work exactly for the reason given, but
I didn't know what else to try.

Now that my only idea is gone, the question for y'all is: how?

=EF=BB=BF(Okay, I actually do have one more idea: construct the insert
valu=
es
list with quote_literal. I'm hoping there's a better way.)=20

Thanks,
Reece

--=20
Reece Hart, http://harts.net/reece/,
GPG:0x25EC91A0


--=20
Sent via pgsql-general mailing list (pgsql-general@[EMAIL PROTECTED]
)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general