[BUGS] Weird misinterpretation of EXECUTE in PL/pgSQL

2007-02-09 Thread Peter Eisentraut
PostgreSQL 8.2.1 This is OK: test=# create function test1() returns int language plpgsql as $$begin execute 'select a, b'; end$$; CREATE FUNCTION test=# select test1(); ERROR: column "a" does not exist LINE 1: select a, b ^ QUERY: select a, b CONTEXT: PL/pgSQL function "test1"

Re: [BUGS] Weird misinterpretation of EXECUTE in PL/pgSQL

2007-02-09 Thread Tom Lane
Peter Eisentraut <[EMAIL PROTECTED]> writes: > In the presence of a comma-separated list of arguments it seems to > intepret the EXECUTE command wildly differently. What is going on here? Basically it evaluates the string-producing argument by sticking SELECT on the front and handing it to the ma

Re: [BUGS] BUG #2393: update fails with unique constraint violation

2007-02-09 Thread Lar
I just wanted to see if there is any plan to develop a solution to this - I still see that there is a todo listed on the postgresql site at http://developer.postgresql.org/cvsweb.cgi/~checkout~/pgsql/doc/src/FAQ/TODO.html ...but it hasn't been visited since September 06. There was a motivation

Re: [BUGS] BUG #2393: update fails with unique constraint violation

2007-02-09 Thread Tom Lane
Lar <[EMAIL PROTECTED]> writes: > I just wanted to see if there is any plan to develop a solution to > this Nothing is likely to happen until someone has a great idea about how to do it without a major performance hit. And you can't have great ideas on a schedule. regard