neilconway opened a new issue, #25704:
URL: https://github.com/apache/datafusion/issues/25704
### Describe the bug
`EliminateCrossJoin` flattens every inner join beneath an inner-join root,
merges all their equijoin keys, and rebuilds the joins using only the root
join's `null_equality`. If a nested inner join carries `NullEqualsNull`, the
rebuilt join silently downgrades it to `NullEqualsNothing`, and NULL keys stop
matching.
### To Reproduce
```
CREATE TABLE t1(a INT, b INT) AS VALUES (NULL, 1), (5, 2);
CREATE TABLE t2(a INT) AS VALUES (NULL), (5);
CREATE TABLE t3(b INT) AS VALUES (1), (2);
-- Correct: 2 rows, NULL matches NULL
SELECT t1.a, t1.b, t2.a
FROM t1 JOIN t2 ON t1.a IS NOT DISTINCT FROM t2.a
ORDER BY t1.b;
-- Wrong: 1 row, the NULL-key row is gone
SELECT t1.a, t1.b, t2.a, t3.b
FROM t1 JOIN t2 ON t1.a IS NOT DISTINCT FROM t2.a
JOIN t3 ON t1.b = t3.b
ORDER BY t1.b;
```
Output:
```
+------+---+------+
| a | b | a |
+------+---+------+
| NULL | 1 | NULL |
| 5 | 2 | 5 |
+------+---+------+
2 row(s) fetched.
Elapsed 0.022 seconds.
+---+---+---+---+
| a | b | a | b |
+---+---+---+---+
| 5 | 2 | 5 | 2 |
+---+---+---+---+
1 row(s) fetched.
Elapsed 0.002 seconds.
```
### Expected behavior
Both queries should return the NULL-key row.
### Additional context
_No response_
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
