While working on a bug in postgres_fdw, I found this:
-> Seq Scan on SmallTable1 A
-> Seq Scan on SmallTable2 B
-> Index Scan using XYIndex on LargeTable C
Index Condition: C.X = A.AID and C.Y = B.BID
This seems odd to me; we would not need the bottom-level Nestloop.
Attached is a small patch for fixing that. I also adjusted the
indentation to the last line "Index Condition: ...", to match others in
the section of Parameterized Paths.
pgsql-hackers by date
|Next:||From: Amit Kapila||Date: 2017-01-23 10:17:36|
|Subject: Re: pageinspect: Hash index support|
|Previous:||From: Etsuro Fujita||Date: 2017-01-23 09:56:07|
|Subject: Re: postgres_fdw bug in 9.6|