Question on explain

[Enrico Pirozzi <sscotty71@xxxxxxxxx>]

Hi all,

I have 2 tables:

table1 (field1,field2,.....) 
table2 (field1,field2,field3,.....)

field1 is an uuid type
field2 is a timestamp with time zone type.

If I execute:

explain (analyze on, timing off)
select B.field3,A.field1,A.field2
FROM table1  A INNER JOIN table2 B
using (field1,field2)

the query plan is


                                                QUERY PLAN                                                 
   
----------------------------------------------------------------------------------------------------------------
 Hash Join  (cost=137324.20..247200.77 rows=1 width=31) (actual rows=1136175 loops=1)
   Hash Cond: ((b.field1 = a.field1) AND (b.field2 = a.field2))
   ->  Seq Scan on table2 b  (cost=0.00..49694.75 rows=1136175 width=31) (actual rows=1136175 loops
=1)
   ->  Hash  (cost=89610.68..89610.68 rows=2287368 width=24) (actual rows=1143684 loops=1)
         Buckets: 65536  Batches: 4  Memory Usage: 15699kB
         ->  Seq Scan on table1 a  (cost=0.00..89610.68 rows=2287368 width=24) (actual rows=1143684 loops=1
)
 Total runtime: 5055.118 ms
(7 rows)

My question is
Why Have I a rows=1 in the explain and rows=1136175 in the explain analyze?

I tried to tune Planner Cost Constants as

default_statistics_target (integer)

and / or

from_collapse_limit (integer)
join_collapse_limit (integer)

but the query plan has been the same.

Can anyone help me ?
Thanks,

Enrico

--
That's one small step for man; one giant leap for mankind

Enrico Pirozzi
Tel.  +39 0861 1855771 - Mob.+39 328 4164437 - Fax  +39 0861 1850310
http://www.pgtraining.com <http://www.pgtraining.com-> -
info@xxxxxxxxxxxxxx
www.enricopirozzi.info - info@xxxxxxxxxxxxxxxxxx
Skype sscotty71 - Gtalk sscotty71@xxxxxxxxx