6

In this answer I explain SQL-89s implicit syntax.
But I noticed different query plans while playing around:

EXPLAIN ANALYZE
  SELECT *
  FROM (values(1)) AS t(x), (values(2)) AS g(y);

                                     QUERY PLAN                                     
------------------------------------------------------------------------------------
 Result  (cost=0.00..0.01 rows=1 width=0) (actual time=0.002..0.002 rows=1 loops=1)
 Planning time: 0.052 ms
 Execution time: 0.020 ms
(3 rows)

Versus this:

EXPLAIN ANALYZE
  SELECT *
  FROM (values(1)) AS t(x)                      
  CROSS JOIN (values(2)) AS g(y);
                                           QUERY PLAN                                           
------------------------------------------------------------------------------------------------
 Subquery Scan on g  (cost=0.00..0.02 rows=1 width=4) (actual time=0.004..0.005 rows=1 loops=1)
   ->  Result  (cost=0.00..0.01 rows=1 width=0) (actual time=0.002..0.002 rows=1 loops=1)
 Planning time: 0.075 ms
 Execution time: 0.027 ms
(4 rows)

Why would one of them show Subquery Scan?
Isn't the implicit syntax just rewritten to be the same as the explicit syntax?

11

Isn't the implicit syntax just rewritten to be the same as the explicit syntax?

Not necessarily. You are building on slightly incorrect assumptions. Like I explained under the referenced question:

Comma-separated items in the FROM list are almost, but not quite identical to explicit CROSS JOIN notation. Explicit joins bind stronger. The query planner has to handle both cases differently in certain cases.

Obviously, the planner is smart enough to be able to treat VALUES expression with a single row expression with a simplified plan. We see a more complex plan for more than one row in the VALUES expression:

EXPLAIN ANALYZE
SELECT *
FROM  (VALUES (1), (2)) t(x)
    , (VALUES (2), (3)) g(y);
 Nested Loop  (cost=0.00..0.11 rows=4 width=8) (actual time=0.059..0.064 rows=4 loops=1)
   ->  Values Scan on "*VALUES*"  (cost=0.00..0.03 rows=2 width=4) (actual time=0.004..0.004 rows=2 loops=1)
   ->  Materialize  (cost=0.00..0.04 rows=2 width=4) (actual time=0.025..0.026 rows=2 loops=2)
         ->  Values Scan on "*VALUES*_1"  (cost=0.00..0.03 rows=2 width=4) (actual time=0.001..0.002 rows=2 loops=1)

Simplifying the query plan is easier for a VALUES expression separated by comma(s). When bound by explicit join(s), Postgres needs to consider join conditions before combining with other comma-separated FROM items. I expect the "subqery scan" we see for this case is a (completely harmless) side effect of the more complex code path for this case.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.