Re: Aggregates push-down to partitions

On 10.11.2017 12:15, Ashutosh Bapat wrote:
> Maybe in this thread[1] your described problem are solved through
>> introducing Parallel Append node?
>>
>> 1.
>> https://www.postgresql.org/message-id/CAJ3gD9dy0K_E8r727heqXoBmWZ83HwLFwdcaSSmBQ1%2BS%2BvRuUQ%40mail.gmail.com
> You may want to review [2] and [3] as well.
>
> [2] https://www.postgresql.org/message-id/9666.1491295317@localhost
> [3] https://www.postgresql.org/message-id/CAM2+6=V64_xhstVHie0Rz=KPEQnLJMZt_e314P0jaT_oJ9MR8A@mail.gmail.com
Thank you very much for this references.
I applied partition-wise-agg-v6 patches and for partitioned tables it
works perfectly:

shard=# explain select count(*) from orders;
                                      QUERY PLAN
---------------------------------------------------------------------------------------
 Finalize Aggregate  (cost=100415.29..100415.30 rows=1 width=8)
   ->  Append  (cost=50207.63..100415.29 rows=2 width=8)
         ->  Partial Aggregate  (cost=50207.63..50207.64 rows=1 width=8)
               ->  Foreign Scan on orders_0 (cost=101.00..50195.13
rows=5000 width=0)
         ->  Partial Aggregate  (cost=50207.63..50207.64 rows=1 width=8)
               ->  Foreign Scan on orders_1 (cost=101.00..50195.13
rows=5000 width=0)
(6 rows)

But I wonder why the same optimization is not applied to normal
inherited table:

shard=# explain select count(*) from base;
                                    QUERY PLAN
----------------------------------------------------------------------------------
 Aggregate  (cost=44087.99..44088.00 rows=1 width=8)
   ->  Append  (cost=0.00..39079.46 rows=2003414 width=0)
         ->  Seq Scan on base  (cost=0.00..0.00 rows=1 width=0)
         ->  Seq Scan on derived1  (cost=0.00..14425.00 rows=1000000
width=0)
         ->  Seq Scan on derived2  (cost=0.00..14425.00 rows=1000000
width=0)
         ->  Foreign Scan on derived_fdw  (cost=100.00..212.39
rows=3413 width=0)
(6 rows)

Are there some principle problems?

--
Konstantin Knizhnik
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company