Konstantin Orlov created IGNITE-23970:
-----------------------------------------

             Summary: Sql. Extend usage of HashJoin
                 Key: IGNITE-23970
                 URL: https://issues.apache.org/jira/browse/IGNITE-23970
             Project: Ignite
          Issue Type: Improvement
          Components: sql
            Reporter: Konstantin Orlov


At the moment, we apply HashJoin only when all predicates involving fields of 
the tables are EQUALITY predicates, but in fact we can take an advantage of 
HashJoin even when some of the predicates are not EQUALITY predicate. Let's 
take a look at example below:

{code:java}
CREATE TABLE t1 (id INT PRIMARY KEY, val INT);
CREATE TABLE t2 (id INT PRIMARY KEY, val INT);

SELECT * FROM t1, t2 WHERE t1.id = t2.id AND t1.val > t2.val;
{code}

We still can use HashJoin to join tables by ID columns with additional post 
filtration {{t1.val > t2.val}}.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to