Re: v9.1, DROP TRIGGER IF EXISTS behaving oddly

On 01/14/2016 01:47 PM, Williamson, Michael wrote:
> Please excuse my mistake.
>
> We were dropping a view that had the trigger on it beforehand, then in
> trying to re-build the entire schema had scripts that attempted to drop
> trigger if they existed before re-creating the view and triggers.  I
> over-sanitized the example I posted and made it unclear.
>
> I did some more testing and DROP TRIGGER IF EXISTS indeed works as
> expected if the table/view is there and the trigger is not there.
>
> As a follow-up, it would be nice if "IF EXISTS" could apply to either
> the trigger or the relation it is applied to, as I don't think a
> trigger can't exist without a relation to apply it to.

That is the part Tom and David where trying to explain to me and I was
not getting, the behavior has changed.

Some testing revealed this:

test=# select version();
                                                              version


----------------------------------------------------------------------------------------------------------------------------------
  PostgreSQL 9.0.20 on i686-pc-linux-gnu, compiled by GCC gcc (SUSE
Linux) 4.8.1 20130909 [gcc-4_8-branch revision 202388], 32-bit
(1 row)


test=# drop trigger if exists test_trg on test_tbl;
ERROR:  relation "test_tbl" does not exist


test=# select version();
                                                            version


-----------------------------------------------------------------------------------------------------------------------------
  PostgreSQL 9.4.5 on i686-pc-linux-gnu, compiled by gcc (SUSE Linux)
4.8.3 20140627 [gcc-4_8-branch revision 212064], 32-bit
(1 row)


test=# drop trigger if exists test_trg on test_tbl;
NOTICE:  relation "test_tbl" does not exist, skipping
DROP TRIGGER

In either case though IF EXISTS did the right thing, so I am wondering
if upgrading to a more recent version of 9.1 will help you.


>
> Michael
>
>



--
Adrian Klaver
adrian.klaver@aklaver.com