ozankabak commented on code in PR #12003:
URL: https://github.com/apache/datafusion/pull/12003#discussion_r1720839499
##########
datafusion/sqllogictest/test_files/joins.slt:
##########
@@ -3898,6 +3898,7 @@ SELECT * FROM (
) as lhs RIGHT JOIN (
SELECT * from right_table_no_nulls
ORDER BY b
+ LIMIT 10
Review Comment:
We can not leave the test as is because the sort in the subquery gets
optimized away (rightfully), breaking the intent of the test. These kind of
tests seems to be originally written with the assumption that the sort wouldn't
be optimized away, and the subquery with ORDER BY was mimicking a table with an
order.
Placing the limit prevents the optimizer from optimizing the sort away,
keeping the intent of the test intact. Keeping the test as is gives us a
degenerate test with our new, more powerful optimizer.
Once we get `WITH ORDER` support for memory tables, we will simply use a
memory table with ordering instead of the subquery and those tests will be
reverted back. In the interim, having the LIMIT is the safest (i.e. least
test-changing) path forward
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org
For additional commands, e-mail: github-h...@datafusion.apache.org
