Re: [HACKERS] Allow GiST opcalsses without compress\decompres functions

Поиск
Список
Период
Сортировка
От Dmitriy Sarafannikov
Тема Re: [HACKERS] Allow GiST opcalsses without compress\decompres functions
Дата
Msg-id 20170914112014.1354.75663.pgcf@coridan.postgresql.org
обсуждение исходный текст
Ответ на Re: [HACKERS] Allow GiST opcalsses without compress\decompres functions  (Andrey Borodin <x4mmm@yandex-team.ru>)
Ответы Re: [HACKERS] Allow GiST opcalsses without compress\decompres functions  (Alexander Korotkov <a.korotkov@postgrespro.ru>)
Список pgsql-hackers
The following review has been posted through the commitfest application:
make installcheck-world:  tested, passed
Implements feature:       tested, passed
Spec compliant:           tested, passed
Documentation:            tested, passed

This is simple and intuitive patch. Code looks pretty clear and well documented.

I think it is good idea to decrease overhead on calling fake compressing/decomressing
functions. This eliminates the need to implement the fetch function in such cases.

I also tried to disable compress and decompress functions in contrib/cube:
diff --git a/contrib/cube/cube--1.2.sql b/contrib/cube/cube--1.2.sql
index dea2614888..26301b71d7 100644
--- a/contrib/cube/cube--1.2.sql
+++ b/contrib/cube/cube--1.2.sql
@@ -370,8 +370,6 @@ CREATE OPERATOR CLASS gist_cube_ops
       FUNCTION        1       g_cube_consistent (internal, cube, smallint, oid, internal),       FUNCTION        2
 g_cube_union (internal, internal),
 
-       FUNCTION        3       g_cube_compress (internal),
-       FUNCTION        4       g_cube_decompress (internal),       FUNCTION        5       g_cube_penalty (internal,
internal,internal),       FUNCTION        6       g_cube_picksplit (internal, internal),       FUNCTION        7
g_cube_same(cube, cube, internal),
 

And it is enables IndexOnlyScan, this is expected behaviour.
+ -- test explain
+ set enable_seqscan to 'off';
+ set enable_bitmapscan to 'off';
+ select count(*) from test_cube where c && '(3000,1000),(0,0)';
+  count
+ -------
+      5
+ (1 row)
+
+ explain analyze select c from test_cube where c && '(3000,1000),(0,0)';
+                                                            QUERY PLAN
+
--------------------------------------------------------------------------------------------------------------------------------
+  Index Only Scan using test_cube_ix on test_cube  (cost=0.15..56.43 rows=16 width=32) (actual time=0.015..0.018
rows=5loops=1)
 
+    Index Cond: (c && '(3000, 1000),(0, 0)'::cube)
+    Heap Fetches: 5
+  Planning time: 0.038 ms
+  Execution time: 0.032 ms
+ (5 rows)

The new status of this patch is: Ready for Committer

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

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Michael Paquier
Дата:
Сообщение: Re: [HACKERS] Setting pd_lower in GIN metapage
Следующее
От: Amit Khandekar
Дата:
Сообщение: Re: [HACKERS] expanding inheritance in partition bound order