Re: [HACKERS] Allow GiST opcalsses without compress\decompresfunctions

Hi, Tom!
> 20 сент. 2017 г., в 8:38, Tom Lane <tgl@sss.pgh.pa.us> написал(а):
>
> Andrey Borodin <x4mmm@yandex-team.ru> writes:
>> [ 0001-Allow-uncompressed-GiST-4.patch ]
>
> Pushed, with a bit more work on the documentation and some minor
> cosmetic changes.
>
> I did not like the fact that the new code paths added by the patch
> were untested, so I went ahead and removed the no-longer-needed
> no-op functions in the core GiST opclasses.  There's still room
> to improve the contrib opclasses, but that's much more tedious
> (you need to write an extension update script) so I didn't feel
> like messing with those now.

I was looking for a way to correctly drop compress\decompress functions from opclasses.
There are two cases: when the functions do something unnecessary (like TOASTing) and when they are just no-ops.

First case occurs, for example, in cube. Please see patch attached. There is no DROP DEFAULT and I'm doing
UPDATE pg_opclass SET opcdefault = FALSE WHERE opcname = 'gist_cube_ops';
I'm not sure it is correct way, but I haven't found any other workaround. Then I just create new default opclass
withoutcompress\decompress. 

Second case, for example in seg. I have tried
alter operator family gist_seg_ops using gist drop function 3 (seg);
It does not work: dependency between opclass and method is DEPENDENCY_INTERNAL and command is refused.
Should I create new opclass as with cube or, may be, I can delete functions from system catalog?

Best regards, Andrey Borodin.


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers