Re: Bad optimizer data for xml (WAS: xml data type implications of no =)

On 09/06/10 15:22, Robert Haas wrote:
> On Thu, Jun 3, 2010 at 7:16 PM, Mark Kirkwood
> <mark(dot)kirkwood(at)catalyst(dot)net(dot)nz> wrote:
>
>> Maybe I gave this guy a bad title - is it a concern that the 'width'
>> estimate is so far off for xml datatypes (because of no = op)? It seemed to
>> me that this could result in some bad plan choices (e.g in subqueries etc).
>>
> It's possible. I don't really see a reason not to add an = operator
> for XML - does anyone else?
>
> It would need to be done by updating src/include/catalog/pg_*.h,
> rather than via SQL, of course.
>
>

Heh, sure should - I merely included the SQL stuff in case anyone else
wanted to reproduce what I was seeing!

Cheers

Mark