Re: Inheritance, unique keys and performance

[Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>]

"Julian Scarfe" <julian(at)avbrief(dot)com> writes:
> From: "Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us>
>> The
>> problem you've got is that you need a nestloop-with-inner-indexscan
>> plan type, where the inner side is an Append group (that is, an
>> inheritance tree) --- and 8.1 doesn't know how to create such a plan.
>> If you can update to 8.2 or later it should get better.

> Isn't test 3 (scalar subqueries) also a nestloop-with-inner-indexscan with 
> an inner Append?

No.  The subquery is planned separately and sees the upper query's
variable as a pseudo-constant Param.  In neither query does the
planner think that a join is happening.

			regards, tom lane