Re: Substantial bloat in postgres_fdw regression test runtime

Ashutosh Bapat <ashutosh(dot)bapat(at)enterprisedb(dot)com> writes:
> On Thu, Nov 3, 2016 at 1:58 PM, Jeevan Chalke
> <jeevan(dot)chalke(at)enterprisedb(dot)com> wrote:
>> Attached patch with test-case modification.

> I verified that this patch indeed bring the time down to 2 to 3
> seconds from 10 seconds.

Thanks for working on this, guys.

> The additional condition t2.c2 = 6 seems to echo the filter t2.c2 = 6
> of aggregate. We wouldn't know which of those actually worked. I
> modified the testcase to use t2.c2 % 6 = 0 instead and keep the filter
> condition intact. This increases the execution time by .2s, which may
> be ok. Let me know what you thing of the attached patch.

Agreed, that seems like a good compromise. Pushed that way.

> Also, please add this to the commitfest, so that it isn't forgotten.

No need.

regards, tom lane