Question

SELECT c.my_type || '-' || cs.my_site, cs.id
FROM table2 cs
JOIN table1 c ON cs.id = c.id
WHERE c.status = 'ACTIVE'

EXPLAIN ANALYZE：

Hash Join  (cost=927.36..66271.66 rows=1141298 width=24) (actual time=9.221..888.461 rows=825843 loops=1)
  Hash Cond: (cs.id = c.id)
  ->  Seq Scan on table2 cs  (cost=0.00..29630.92 rows=1652792 width=21) (actual time=0.006..320.526 rows=1668338 loops=1)
  ->  Hash  (cost=757.70..757.70 rows=13573 width=7) (actual time=8.995..8.995 rows=13044 loops=1)
        ->  Seq Scan on tabl1 c  (cost=0.00..757.70 rows=13573 width=7) (actual time=0.005..6.005 rows=13044 loops=1)
              Filter: ((status)::text = 'ACTIVE'::text)
Total runtime: 948.947 ms

vs

select c.my_type || '-' || cs.my_site, cs.id 
from (SELECT id, my_site FROM table1 WHERE status = 'ACTIVE') as c
join table2 cs on cs.id = c.id

EXPLAIN ANALYZE

Hash Join  (cost=927.36..66271.66 rows=1141298 width=24) (actual time=8.927..894.598 rows=821287 loops=1)
  Hash Cond: (cs.id = c.id)
  ->  Seq Scan on table2 cs  (cost=0.00..29630.92 rows=1652792 width=21) (actual time=0.005..331.769 rows=1663782 loops=1)
  ->  Hash  (cost=757.70..757.70 rows=13573 width=7) (actual time=8.706..8.706 rows=13044 loops=1)
        ->  Seq Scan on table1  (cost=0.00..757.70 rows=13573 width=7) (actual time=0.005..6.044 rows=13044 loops=1)
              Filter: ((status)::text = 'ACTIVE'::text)
Total runtime: 954.941 ms

我希望第二个查询比第一个查询相对快..但它似乎不是。请记住，这些是非常大的表格。

Postgresql查询优化连接

0 个答案: