the output is: pba=# EXPLAIN (ANALYZE, TIMING, BUFFERS) SELECT idprodxintegrar FROM _gc_tb a LEFT join _gc_cat b on ( b.arama <@ a.arama and a.arama <@ b.arama ) ; QUERY PLAN --------------------------------------------------------------------------------------------------------- Nested Loop Left Join (cost=0.00..0.01 rows=1 width=4) (actual time=0.003..0.003 rows=0 loops=1) Join Filter: ((b.arama <@ a.arama) AND (a.arama <@ b.arama)) -> Seq Scan on _gc_tb a (cost=0.00..0.00 rows=1 width=66) (actual time=0.002..0.002 rows=0 loops=1) -> Seq Scan on _gc_cat b (cost=0.00..0.00 rows=1 width=70) (never executed) Planning time: 0.206 ms Execution time: 0.074 ms (6 filas)
2016-03-04 15:01 GMT-06:00 David G. Johnston <david.g.johns...@gmail.com>: > On Fri, Mar 4, 2016 at 1:52 PM, Felipe de Jesús Molina Bravo < > fjmolinabr...@gmail.com> wrote: > >> >> >>>> Now i execute the same in pgsql 9.4.5 and all is fine!!! >>>> >>>> The EXPLAINs are: >>>> >>>> - pgsql 9.5.1: >>>> >>>> Nested Loop Left Join (cost=0.03..492944.81 rows=276095 width=4) >>>> -> Seq Scan on _gc_tb a (cost=0.00..3321.30 rows=120130 width=66) >>>> -> Bitmap Heap Scan on _gc_cat b (cost=0.03..4.06 rows=2 >>>> width=70) >>>> Recheck Cond: ((arama <@ a.arama) AND (a.arama <@ arama)) >>>> -> Bitmap Index Scan on _gc_cat_arama_gin >>>> (cost=0.00..0.03 rows=2 width=0) >>>> Index Cond: ((arama <@ a.arama) AND >>>> (a.arama <@ arama)) >>>> >>>> >>>> - pgsql 9.4.5: >>>> Nested Loop Left Join (cost=0.03..492944.81 rows=276095 width=4) >>>> -> Seq Scan on _gc_tb a (cost=0.00..3321.30 rows=120130 width=66) >>>> -> Bitmap Heap Scan on _gc_cat b (cost=0.03..4.06 rows=2 >>>> width=70) >>>> Recheck Cond: ((arama <@ a.arama) AND (a.arama <@ arama)) >>>> -> Bitmap Index Scan on _gc_cat_arama_gin >>>> (cost=0.00..0.03 rows=2 width=0) >>>> Index Cond: ((arama <@ a.arama) >>>> AND >>>> (a.arama <@ arama)) >>>> >>> >>> The above are exactly the same, so if they are indeed from the different >>> versions I do not see an issue. The question to ask here is whether the >>> above are actually from the different Postgres instances? >>> >>> >> yes these are differents >> > > It would be nice to see the output of "EXPLAIN (ANALYZE, TIMING, > BUFFERS)" so real timings can be observed. > > > David J. > > >