Which SQL query is faster? Filter on Join criteria or Where clause?

Question

Compare these 2 queries. Is it faster to put the filter on the join criteria or in the WHERE clause. I have always felt that it is faster on the join criteria b

Answer 1

In postgresql they are the same. We know this because if you do explain analyze on each of the queries, the plan comes out to be the same. Take this example:

# explain analyze select e.* from event e join result r on e.id = r.event_id and r.team_2_score=24;

                                                  QUERY PLAN                                                   
---------------------------------------------------------------------------------------------------------------
 Hash Join  (cost=27.09..38.22 rows=7 width=899) (actual time=0.045..0.047 rows=1 loops=1)
   Hash Cond: (e.id = r.event_id)
   ->  Seq Scan on event e  (cost=0.00..10.80 rows=80 width=899) (actual time=0.009..0.010 rows=2 loops=1)
   ->  Hash  (cost=27.00..27.00 rows=7 width=8) (actual time=0.017..0.017 rows=1 loops=1)
         Buckets: 1024  Batches: 1  Memory Usage: 9kB
         ->  Seq Scan on result r  (cost=0.00..27.00 rows=7 width=8) (actual time=0.006..0.008 rows=1 loops=1)
               Filter: (team_2_score = 24)
               Rows Removed by Filter: 1
 Planning time: 0.182 ms
 Execution time: 0.101 ms
(10 rows)

# explain analyze select e.* from event e join result r on e.id = r.event_id where r.team_2_score=24;
                                                  QUERY PLAN                                                   
---------------------------------------------------------------------------------------------------------------
 Hash Join  (cost=27.09..38.22 rows=7 width=899) (actual time=0.027..0.029 rows=1 loops=1)
   Hash Cond: (e.id = r.event_id)
   ->  Seq Scan on event e  (cost=0.00..10.80 rows=80 width=899) (actual time=0.010..0.011 rows=2 loops=1)
   ->  Hash  (cost=27.00..27.00 rows=7 width=8) (actual time=0.010..0.010 rows=1 loops=1)
         Buckets: 1024  Batches: 1  Memory Usage: 9kB
         ->  Seq Scan on result r  (cost=0.00..27.00 rows=7 width=8) (actual time=0.006..0.007 rows=1 loops=1)
               Filter: (team_2_score = 24)
               Rows Removed by Filter: 1
 Planning time: 0.140 ms
 Execution time: 0.058 ms
(10 rows)

They both have the same min and max cost as well as the same query plan. Also, notice that even in the top query the team_score_2 gets applied as a 'Filter'.