Subject: Re: Delete query takes exorbitant amount of time
From: Simon Riggs <simon@2ndquadrant.com>
To: Tom Lane <tgl@sss.pgh.pa.us>, Karim Nassar <karim.nassar@NAU.EDU>
Cc: Christopher Kings-Lynne <chriskl@familyhealth.com.au>,
	pgsql-performance@postgresql.org
In-Reply-To: <20845.1111733900@sss.pgh.pa.us>
References: <1111709457.9085.127.camel@k2.cet.nau.edu>
	<2630.1111711927@sss.pgh.pa.us>
	<1111715131.9085.174.camel@k2.cet.nau.edu>
	<3750.1111715318@sss.pgh.pa.us> <4243879B.10200@familyhealth.com.au>
	<20845.1111733900@sss.pgh.pa.us>
Content-Type: text/plain
Organization: 2nd Quadrant
Date: Fri, 25 Mar 2005 15:10:48 +0000
Message-Id: <1111763448.11750.730.camel@localhost.localdomain>
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit

On Fri, 2005-03-25 at 01:58 -0500, Tom Lane wrote:
> Christopher Kings-Lynne <chriskl@familyhealth.com.au> writes:
> >> In that case there's a datatype mismatch between the referencing and
> >> referenced columns, which prevents the index from being used for the
> >> FK check.
> 
> > Is creating such a foreign key a WARNING yet?
> 
> I believe so as of 8.0.  It's a bit tricky since 8.0 does allow some
> cross-type cases to be indexed, but IIRC we have a test that understands
> about that...

src/backend/commands/tablecmds.c, line 3966 in CVSTIP
/*
 * Check that the found operator is compatible with the PK index,
 * and generate a warning if not, since otherwise costly seqscans
 * will be incurred to check FK validity.
*/
if (!op_in_opclass(oprid(o), opclasses[i]))
   ereport(WARNING,
	(errmsg("foreign key constraint \"%s\" "
		"will require costly sequential scans",
		fkconstraint->constr_name),
	 errdetail("Key columns \"%s\" and \"%s\" "
	   	"are of different types: %s and %s.",
		 strVal(list_nth(fkconstraint->fk_attrs, i)),
		 strVal(list_nth(fkconstraint->pk_attrs, i)),
		   format_type_be(fktypoid[i]),
		   format_type_be(pktypoid[i]))));

So, yes to the WARNING. Not sure about the cross-type cases...

Karim: Did this happen? If not, can you drop and re-create and confirm
that you get the WARNING? If not, we have problems.

I vote to make this an ERROR in 8.1 - I see little benefit in allowing
this situation to continue. If users do create a FK like this, it just
becomes another performance problem on list...

Best Regards, Simon Riggs