On Tue, Jul 12, 2011 at 12:10 PM, Shianmiin <Shianmiin(at)gmail(dot)com> wrote:
>
> Merlin Moncure-2 wrote:
>>
>> One proposed solution is to cache plpgsql plans around the search path.
>>
>
> I like the proposed solution, since search_path plays a part when generating
> plpgsql plan, it make sense to be part of the cache.
>
>
> Merlin Moncure-2 wrote:
>>
>> *) use sql functions for portions that float across schemas
>>
>
> Just to clarify, does this mean the sql functions doesn't cache plans like
> plpgsql functions do?
correct. so you could wrap schema dependent bits inside set returning
sql functions.
merlin